JP6501398B2 - Concept processing device and program - Google Patents

Description

  The present invention relates to a concept processing device and program for processing a concept. The present invention particularly relates to a concept processing apparatus and program for determining the degree of association between concepts.

  It is expected that more advanced processing can be performed by utilizing relationships between concepts in information retrieval processing, knowledge processing using natural language, and the like. It is also expected to be able to efficiently collect the relationships between concepts without human intervention.

  As information representing a relation between concepts, there can be considered a relation name between concepts, and a numerical value representing the strength of a relation between concepts. Among these, if the strength of association between concepts can be collected accurately and efficiently, it can be expected to contribute to the advancement of the information search processing and the knowledge processing described above.

  Non-Patent Document 1 describes a technique for generating a large-scale similar word list using the results of large-scale probabilistic clustering of dependencies. This technology presupposes a large-scale analysis of linguistic resources. And this technology defines a context with words appearing around a word or words in a dependency relationship based on a distribution hypothesis that words similar in meaning appear in similar context, Analyze co-occurrence frequency etc.

  In Non-Patent Document 2, it is described that a database in which the degree of association between words is calculated is published by analyzing a large amount of sentences and finding the frequency of appearing in common in the same sentence. As co-occurrence scores, three types of co-occurrence frequency, Dice coefficient, and discounting mutual information amount are used.

Yuichi Kazama, Stijn De Saeger, Kentaro Torizawa, Maki Murata, "Creating a Large-scale Similar Word List Using Probabilistic Clustering of Dependencies," Proceedings of the 15th Annual Conference of the Association for Language Processing, March 2009 , P. 84-87, [online], [search on April 1, 2015], Internet <URL: http://www.anlp.jp/proceedings/annual_meeting/2009/pdf_dir/C1-6.pdf> ALAGIN language resource / voice resource site, “(A-5) Word co-occurrence frequency database”, [online], [April 10, 2015 search], Internet <URL: https: //alaginrc.nict.go. jp / resources / nict-resource / li-info / li-outline.html>

  In the method according to Non-Patent Document 1, for example, as in the concept pair “paeria” and “clam”, it is not possible to set a high degree of association with a pair having a food-material relationship. Therefore, good performance can not be obtained even if the method of Non-Patent Document 1 is applied as it is to the degree of association in the concept map.

  Also, although the method according to Non-Patent Document 2 is effective for the relationship of concept pairs such as cooking and materials, for example, similar usage such as concept pairs of “soccer” and “baseball” It is not possible to give a high value to word pairs that are used in pairs but rarely used in the same context. Therefore, it is difficult to apply this method as it is as the degree of association in the concept map.

  In addition, both the method according to Non-Patent Document 1 and the method according to Non-Patent Document 2 have to analyze large-scale language resources.

  The present invention has been made based on the recognition of these problems, and a concept processing device and program capable of obtaining an index (degree of association) indicating the strength of association between concepts based on a concept map. It is intended to be provided.

  In order to solve the above problems, a concept processing device according to an aspect of the present invention reads concept map data representing the presence or absence of a relationship between concepts, and the number of links representing a relationship between a certain concept and another concept is A conceptual weight calculation unit that calculates the value of the weight of the concept so as to increase the value of the concept as the number decreases, and a path formed by connecting a certain concept and another concept by the link in the conceptual map data And a path weight calculation unit that calculates the value of the weight of the path according to the value of the weight of the concept for the concept included on the path, and the above for the path connecting between a certain concept and another concept It is characterized by comprising a degree-of-association calculation unit that calculates the degree of association between the concept and the other concept according to the value of the path weight.

  Further, according to one aspect of the present invention, in the above-described concept processing device, the path weight calculation unit obtains a harmonic mean or a geometric mean of the weight values of the concept for each concept included on the path. The value of the weight of the path is calculated.

  Further, according to one aspect of the present invention, in the above-described concept processing device, the concept weight calculation unit calculates the logarithm of the reciprocal of the number of links between the concept and the other concept. It is characterized in that a weight value is calculated.

  Further, according to one aspect of the present invention, in the above-mentioned concept processing device, when there is a plurality of paths formed by connecting the certain concept and the other concept by the link, the relevance degree calculation unit Calculating the degree of association between the concept and the other concept by multiplying the number of the plurality of paths by the value of the weight of the largest one of the plurality of path weights, or Alternatively, the sum of weights of the plurality of paths is calculated as the degree of association.

  Further, according to one aspect of the present invention, in the above-described concept processing device, a similarity degree acquisition unit that acquires similarity degree data representing similarity between concepts is provided, and the path weight calculation unit is configured to It is characterized in that the similarity of the pair of concepts included in the upper link is acquired from the similarity data, and the value of the path weight is calculated according to the acquired similarity as well.

  Further, one aspect of the present invention is a program for causing a computer to function as any one of the concept processing devices described above.

  According to the present invention, the degree of association between concepts can be determined based on the concept map.

FIG. 2 is a block diagram showing a schematic functional configuration of a concept processing device according to a first embodiment of the present invention. It is a schematic diagram showing an example of the concept map which the concept map storage part by the embodiment holds. It is the schematic for demonstrating the weight of the concept calculated in the same embodiment, and shows a part of concept map. FIG. 14 is a schematic view showing a configuration of data placed in a storage area in a work storage unit according to the embodiment and for holding concept weights and the like. It is the schematic for demonstrating the weight of the path calculated in the same embodiment, and shows a part of concept map. FIG. 17 is a schematic view showing a means for storing information to be temporarily stored when searching for a conceptual map by the path weight calculation unit, which is data placed in a storage area in a work storage unit according to the same embodiment. FIG. 17 is a schematic view showing a means for storing information related to a path to each concept when the path weight calculation unit searches for a concept map, which is data placed in a storage area in a work storage unit according to the same embodiment. It is the 1st of the flowchart which shows the procedure of a process of the concept processing apparatus by the embodiment. It is the 2nd of the flowchart which shows the procedure of a process of the concept processing apparatus by the same embodiment. FIG. 5 is a schematic view showing an example of values held by the storage area shown in FIG. 4 which are data used in the embodiment. FIG. 7 is a schematic view showing an example of transition of values held by the storage area shown in FIG. 6, which is data used in the embodiment. FIG. 8 is a schematic view showing an example (part 1) of transition of values held by the storage area shown in FIG. 7, which is data used in the first embodiment. FIG. 8 is a schematic view showing an example (No. 2) of transition of values held by the storage area shown in FIG. 7, which is data used in the first embodiment. FIG. 8 is a schematic view showing an example (No. 3) of transition of values held by the storage area shown in FIG. 7, which is data used in the first embodiment. It is a block diagram showing a schematic functional configuration of a concept processing device according to a second embodiment of the present invention. It is the schematic which shows the structure of the similarity data which represent the similarity between the concepts which the similarity acquisition part in the embodiment acquires from the exterior.

Next, embodiments of the present invention will be described with reference to the drawings.
First Embodiment
FIG. 1 is a block diagram showing a schematic functional configuration of the conceptual processing device according to the first embodiment. As illustrated, the concept processing device 1 includes a concept weight calculation unit 12, a path weight calculation unit 14, a degree of association calculation unit 16, a parameter setting unit 18, a concept map storage unit 20, and a work storage unit 22. , And an association degree storage unit 30. These units are realized using, for example, an electronic circuit (including a computer). The concept map storage unit 20, the work storage unit 22, and the association degree storage unit 30 are realized by using, for example, a magnetic hard disk drive or a semiconductor memory as a unit for storing information.

The concept weight calculation unit 12 reads concept map data representing the presence or absence of a relationship between concepts from the concept map storage unit 20, and the smaller the number of links representing the relationship between a certain concept and another concept, the larger the relevant concept is. Calculate the value of the weight of the concept that becomes a value.
Also, the concept weight calculator 12 calculates the value of the weight of the concept by calculating the logarithm of the reciprocal of the number of links between the concept and the other concept.

For the path formed by connecting a certain concept and another concept by the link in the concept map data, the path weight calculation unit 14 corresponds to the value of the weight of the concept for the concept included in the path. Calculate the value of the path weight
Further, the path weight calculation unit 14 calculates the value of the weight of the path by taking the harmonic mean or the geometric mean of the value of the weight of the concept regarding each concept included in the path.
In addition, as a specific procedure, the path weight calculation unit 14 starts with the concept of the calculation target, and performs the search processing for searching the conceptual map data, and reaches the concept of the calculation target and the search processing. Calculating the value of the weight of the path for a path between the concept to be calculated, and the value of the weight of the path calculated for the path connecting the concept to be calculated and the reached concept If the calculated path weight value is larger than the path weight value already calculated for the path, the value for the path with the largest weight between the concept to be calculated and the reached concept is calculated. Write to the storage means as the value of.

The degree-of-association calculation unit 16 calculates the degree of association between the concept and the other concept in accordance with the value of the weight of the path connecting a certain concept and the other concept.
Further, when there are a plurality of paths formed by connecting the certain concept and the other concept by the link, the degree-of-relevance calculation unit 16 determines the number of the plurality of paths. The degree of association between the concept and the other concept is calculated by multiplying the value of the largest one of the path weights among the value of the weight.

  The parameter setting unit 18 sets parameters stored in the work storage unit 22. Specifically, the parameter setting unit 18 appropriately sets the value of a parameter MAX_Distance described later.

  The concept map storage unit 20 stores concept map data representing the presence or absence of a relationship between concepts. The details of the conceptual map data will be described later.

  A work storage unit 22 (work memory; sometimes referred to simply as “memory” below) is a storage means and has an area for storing information for work being processed. Further, the work storage unit 22 has an area for storing processing parameters.

  The association degree storage unit 30 stores the association degree between the concepts calculated by the association degree calculation unit 16. The data of the degree of association stored in the association degree storage unit 30 can be referred to as appropriate from another device or the like.

  Next, data of the conceptual map stored in the conceptual map storage unit 20 will be described. The concept map is data indicating whether or not there is a relationship between a certain concept and another concept. One form representing a conceptual map is a graph composed of nodes and links. Each concept corresponds to a node in this graph. Also, the relationships between concepts correspond to links in the graph. There is either a relationship or no relationship between one concept and another. When there is a relationship between concepts, there is a link between the concepts. If there is no relationship between concepts, there is no link between those concepts. In addition, as long as the information showing the presence or absence between a concept and a concept is included, the form of the data of a concept map may be except the thing demonstrated here. In addition, as data of the conceptual map used here, the conceptual map already constructed may be used as it is, or a conceptual map may be newly constructed. One example of how to construct a concept map is to analyze the concepts that appear in the vast amount of language resources. The language resources used here are, for example, web texts that can be collected via the Internet. Also, an analysis method of appearing concepts is, for example, analyzing the appearance of a plurality of concepts in a predetermined syntax, or analyzing a plurality of concepts appearing in the same context. Also, if the data downloadable from the website directly represents the relationship between a plurality of concepts, the relationship obtained there may be used directly. For example, encyclopedia sites such as Wikipedia (URL https://ja.wikipedia.org/) provide relationships between concepts on downloadable mobile phones. An example of a conceptual map is described next.

  FIG. 2 is a schematic view showing an example of the concept map held by the concept map storage unit 20. As shown in FIG. The actual concept map is very large data, and the number of concepts included in the concept map is hundreds of thousands, millions or more. However, in the figure, in order to simplify the explanation, a conceptual map consisting of seven concepts is illustrated. As shown, the conceptual map is represented by an undirected graph. And each node of the graph corresponds to one concept. Also, each link in the graph represents a direct relationship between concepts. There are seven concepts included in the illustrated concept map: "illness", "high blood pressure", "adult disease", "diabetes", "cold", "fever", and "influenza". In addition, when the links included in the conceptual map illustrated are represented by pairs of concepts connected by the links, "disease-high blood pressure", "disease-adult disease", "disease-diabetes", "hypertension-adult disease" , "Diabetes-adult disease", "sickness-cold", "sickness-influenza", "cold-fever", "influenza-fever" nine.

  The conceptual map storage unit 20 holds conceptual map information in a data format that is easy to process. As an example, the concept map storage unit 20 is table data that associates a node identifier with a concept represented by that node, and a table data that associates a link identifier with a pair of two node identifiers connected by the link. Holds conceptual map information. In addition, you may make it apply not only this example but the data format generally used in order to express a graph.

Next, the degree of association between concepts calculated by the concept processing device 1 will be described.
In the present embodiment, the degree of association between concepts is calculated according to how the concepts in the concept map are linked. In the processing by the concept processing device 1, the concepts correspond to the nodes on the graph as described above. Also, there are cases where concepts are connected by links and when they are not. One link connects adjacent concepts, and connects only the adjacent concepts. Concepts can be linked together when they can be reached by traversing the concept map from one concept to another. The condition that the concepts are in a connected state is that one concept can be reached from the other concept by tracing one or more links. The route followed at that time is called a path here. A path consists of one or more links. In the case of multiple links, a path is formed by connecting the links in series. When a concept and another concept are in a connected state, the path connecting the two concepts is not limited to one. You may also be able to reach through several different paths. Note that one pass does not pass the same concept more than once. That is, the path does not form a closed path on the way or return. In the present embodiment, based on the weight of a concept, the weight of a path including the concept is calculated. Also, the degree of association between concepts is calculated based on the weight of the path. The degree of association between concepts also depends on the number of paths connecting the concepts.

  First, the concept processing device 1 calculates the weight of each concept included in the concept map. The weight of a concept is a numerical value representing the importance of the concept in calculating the degree of association. The weight of a concept is calculated according to the number of links leading from the concept to another concept. The smaller the number of links, the larger the value of the weight of the concept. This is based on the premise that the concept is more valuable if the number of links leading from a concept is smaller. More specifically, the weight of a concept is calculated by the logarithm of the reciprocal of the number of links from that concept to other concepts. However, in order to normalize the value of the concept weight, the logarithm of the value obtained by dividing the total number of concepts on the concept map by the number of links described above is taken as the weight of the concept. The base of the logarithm may be determined as appropriate.

  Next, the concept processing device 1 calculates path weights. According to the above definition, there are multiple concepts on the path. Both ends of the path are also conceptual respectively. The weight of a path is calculated by the weight of each concept included in the path. More specifically, the weight of a path is calculated by some mean of the weight of each concept that the path contains. Here, some mean is, for example, a simple arithmetic mean, a harmonic mean, a geometric mean, or the like. More specifically, the concept processing device 1 calculates the value of the harmonic mean of the weight of each concept as the value of the weight of the path. This is because experiments using actual conceptual maps have shown that it is possible to calculate the most preferable path weight by using the harmonic mean.

  Next, the concept processing device 1 calculates the degree of association between concepts based on the value of the path weight. The degree of association between concepts according to this embodiment is calculated as the product of the value of the largest weight among the weights of paths between the concepts and the number of paths existing between the concepts.

  A procedure for calculating the degree of association between concepts with a realistic amount of calculation will be described later. In the procedure, in order to search the concept map with a reasonable amount of calculation, an upper limit is set on the distance to be searched (= 1 number of serial links included in one path). The upper limit value can be set as a parameter. Note that the degree of association between concepts when this procedure is used is directed. That is, the value of the degree of association from concept A to concept B may be different from the value of the degree of association from the opposite concept B to concept A.

In summary, it is as follows.
As described above, the degree of association between concepts calculated by the concept processing device 1 is obtained according to the number of paths existing between the concepts. That is, as the number of paths is larger, the value of the degree of association is larger. This is consistent with the assumption that concepts connected by various paths are highly related to each other.
Further, the degree of association between the concepts calculated by the concept processing device 1 is obtained according to (the maximum value of) the weight of the path. That is, as the maximum value of the path weight is larger, the value of the degree of association is larger. This is consistent with the assumption that concepts connected by paths of high importance are highly related to each other.
Further, the weight of the path calculated by the concept processing device 1 is obtained according to the weight of the concept included in the path. The larger the concept weight, the larger the path weight. It is preferable to use a harmonic mean of concept weights as the path weight, but a geometric mean of concept weights may be used.

Here, the weight of the concept calculated by the concept weight calculator 12 will be described.
FIG. 3 is a schematic diagram for explaining concept weights and shows a part of a concept map. In the figure, the circles labeled "A", "C", "D", and "E" are nodes corresponding to the concept. Also, the lines extending from the nodes are links between the nodes, and represent direct relationships between concepts. Although only a part of the concept map is shown in the figure, the actual concept map extends outside the range shown in the figure.

  Although only a part of the conceptual map is shown in the figure, the number of nodes (the number of concepts) of the entire conceptual map is N. N is a positive integer, for example, N = 10000. In order to explain the weight of the concept, attention is focused on the node labeled “C” shown in the figure as an example. Direct links to nodes "A", "D", and "E" are connected from this node "C". Direct links from nodes "C" to nodes other than nodes "A", "D", and "E" are not connected. That is, the number of links from this node "C" directly to other nodes is three.

The weight of the concept in this embodiment is calculated by the following equation.
Concept weight = log (N / n)
Here, N is the total number of concepts included in the entire concept map as described above. Also, n is the number of concepts directly connected to the concept. In addition, although the base of a logarithm may be determined suitably, let the base be 10 here. In the illustrated example, the number of other nodes directly connected to the node “C” is 3 (n = 3), and the total number of nodes included in the conceptual map is 10000 (N = 10000). Therefore, according to the above equation, the weight of the concept corresponding to the node "C" is calculated as follows.
Weight of concept of node "C" = log ₁₀ (10000/3) = 3.52
However, rounding is performed by rounding off the third decimal place.

  FIG. 4 is a schematic diagram showing the configuration of data for holding concept weights. This data is provided in a predetermined area of the work storage unit 22. The data shown in the figure is tabular data, and has items of node ID, concept (concept corresponding to node), connected concept, and weight. One row of this table corresponds to one concept, ie one node. The node ID is an ID that uniquely identifies a node in the conceptual map. In the illustrated example, an integer value is used as the node ID. The item of concept stores words that express the concept. The connected concept item stores a list of node IDs of concepts connected by a direct link from the concept. The item of weight stores the value of weight calculated for the concept.

  The figure shows a part of the values in the entire table. The node ID is in the range of 98 to 102. Also, the concept of each of these nodes is A, B, C, D, E. Here, focusing on the concept C (the node ID is 100), the node IDs of the concept connected from the concept C are three of 98, 101 and 102. That is, the concept C is directly linked to the concepts A, D, and E (as illustrated in FIG. 3). Also, as the weight of concept C, a value of 3.52 is stored.

Here, the weight of the path calculated by the path weight calculation unit 14 will be described.
FIG. 5 is a schematic diagram for explaining path weights and shows a part of a conceptual map. In the figure, the circles labeled "1", "2", "3", "4" and "5" are nodes corresponding to the concept, respectively. Also, the lines extending from the nodes are links between the nodes, and represent direct relationships between concepts. Although only a part of the concept map is shown in the figure, the actual concept map extends outside the range shown in the figure.

  A path is a route from a concept to a concept. For example, in the figure, there are at least two paths from node “1” to node “5”. The first path is denoted as "1-2-3-5" as a row of nodes. The second path is similarly represented as "1-2-4-5" as a row of nodes. Thus, a path can be represented by arranging the nodes on the path in a row, including the nodes at both ends. And in this embodiment, the weight of a path is defined as the harmonic mean of the weight of each node (including the node of both ends) on the path.

In general, the harmonic mean of _M numerical values w ₁ , w ₂ ,..., W _M is expressed by the following equation. That is, M / (1 / w ₁ + 1 / w ₂ +... + 1 / w _M ).

Therefore, the node "1" weights from up to "5" (i.e., the weight of the concept) is already calculation, their respective _{w n1, w n2, ···,} when the _{w nM,} first the The path weight and the second path weight are respectively expressed by the following equations.
(Weight of first path) = 4 / (1 / w _n1 + 1 / w _n2 + 1 / w _n3 + 1 / w _n5 )
(Weight of second path) = 4 / (1 / w _n1 + 1 / w _n2 + 1 / w _n4 + 1 / w _n5 )

  FIG. 6 is a schematic view showing a part of the configuration of data held by the work storage unit 22. As shown in FIG. The data shown in the figure is mainly means for storing information to be temporarily stored when the path weight calculation unit 14 searches for a concept map while referring to the concept map storage unit 20. The meaning of each data item included in the figure is as follows.

The data item "concept of calculation object" is a concept selected at that time in the concept map, and is a concept for which the degree of association is to be calculated. For example, when the concept to be calculated is "disease" in the concept map shown in FIG. 2, the concept processing device 1 calculates the degree of association from this concept "disease" to each other concept. In addition, the concept of calculation object may be called "the concept A" for convenience below.
The data item “array of starting point concepts” is an array having, as an element, a concept that is a starting point when searching for a path of a predetermined distance (distance of values held by the variable D) from a concept to be calculated.
The data item “the number of concepts traced” is data that holds the distance from the concept to be calculated at that time, and corresponds to the variable D described above.

The data item "concept of origin" is the concept currently selected as the origin. The data item "concept of origin" stores one concept selected from the array of origin concepts described above. Note that the concept of the starting point may be hereinafter referred to as "concept C" for the sake of convenience.
The data item "arrangement of adjacent concepts" is an array having elements adjacent to the concept of the origin described above. Here, the concept "adjacent" means that a certain concept and one other concept are connected via a single link. For example, in the example of the concept map shown in FIG. 2, the concept “disease” and the concept “hypertension” are concepts adjacent to each other. Moreover, the concept "fever" and the concept "cold" are concepts adjacent to each other. Moreover, since the concept "diabetes" and the concept "cold" are not directly linked by a link, they are not adjacent to each other. The adjacent concept (that is, one element of the array of the adjacent concept) may be hereinafter referred to as “concept B” for the sake of convenience.

  A usage example of each data area shown in FIG. 6 will be described later as a processing example.

  FIG. 7 is a schematic view showing a part of the configuration of data held by the work storage unit 22. As shown in FIG. The data shown in the figure mainly stores numerical values and the like for calculating the degree of association of each concept when the path weight calculation unit 14 searches for a concept map while referring to the concept map storage unit 20. is there. Specifically, the path weight calculation unit 14 sets the weight between the concept (concept A) to be calculated and each concept (especially when there are a plurality of paths) according to the processing procedure described later. Update information about the path that becomes the largest one by one. Note that this data is tabular data, and each row of this table corresponds to one concept. The meaning of each data item included in the figure is as follows.

  The data item "concept" stores data of words representing the concept. For example, in the example of the conceptual map shown in FIG. 2, this data item stores information such as “hypertension” or “adult disease” that uniquely identifies the concept. Further, speaking in association with the processing procedure of the path weight calculation unit 14, the concept represented by this data item corresponds to the above-mentioned "concept B".

The data item “MAX_Weight” holds the value of the largest weight among the weights of the paths from the concept to be calculated (concept A) to the concept (concept B). Generally, when the concept A and the concept B are connected within a predetermined distance (MAX_Distance), the number of paths (traceable routes) between both concepts is one or more. The path weight calculation unit 14 updates the data item “MAX_Weight” with the largest value among the path weights already calculated among these paths.
The data item "MAX_Path" stores the information of the path corresponding to the above "MAX_Weight". That is, the data item “MAX_Path” is for holding information of a path between the concept A and the concept B which has the largest calculated weight. Specifically, the data item "MAX_Path" holds a sequence of concepts (including the concepts at both ends of the path) included on the path as information on the path. Such information of the concept column included on the path uniquely identifies the path on the concept map.

The data item "NUM_Path" stores the number of paths from concept A to concept B. When the path weight calculation unit 14 searches for a conceptual map, the value of NUM_Path is calculated as follows. That is, let NUM_Path _{D = n} (A, B) be the number of paths from concept A to concept B when D = n (where 0 ≦ n <MAX_Distance−1). Further, let NUM_Path _{D = n} (A, C) be the number of paths from concept A to concept C when _{D = n} . Here, NUM_Path _{D = n} (A, B) does not include the number of paths passing through the concept C. Then, when D = n + 1 is searched for a direct link path from concept C to concept B (that is, a path from concept A to concept B via concept C is searched at D = n + 1), The following equation holds.

NUM_Path _{D = n + 1} (A, B) = (NUM_Path _{D = n} (A, B)) + (NUM_Path _{D = n} (A, C))

Next, the procedure of processing by the conceptual processing device 1 will be described.
8 and 9 are flowcharts showing the procedure of processing by the concept processing device 1. Both figures are connected to each other by connectors of the flow chart. The procedure will be described below along the flowchart.

  First, in step S10 (FIG. 8), the parameter setting unit 18 assigns the maximum distance (MAX_Distance) as the value of the parameter. Specifically, the parameter setting unit 18 writes the set value of MAX_Distance in a predetermined area in the work storage unit 22. The value of MAX_Distance given is a positive integer. The value of MAX_Distance defines the maximum distance for traversing from concept to concept in the concept map. For example, when a certain concept and a certain concept are connected by a direct link, the distance between the relevant concepts according to the route is 1. Also, when a certain concept and a certain concept are indirectly connected by a link on the way through only one intermediate concept, the distance between the concepts concerned by the route is two. That is, when a certain concept and a certain concept are indirectly connected by a link through one or more intermediate concepts, the number of links when following the route is the distance between the relevant concepts.

  Next, in step S20, the concept weight calculation unit 12 creates a list of concepts in which the concepts connected with the concepts are directly linked to the concepts included in the concept map. The concept weight calculation unit 12 writes the list of the concepts in the column of connected concepts in the table shown in FIG. Also, the concept weight calculation unit 12 calculates the weight of each concept based on the list of connected concepts. The calculation method is based on the formula of log (N / n) as described above. The concept weight calculator 12 writes the values calculated for each concept in the column of weight in the table shown in FIG.

  The processes from the next step S30 to S90 and S100 are processes in which the path weight calculation unit 14 calculates and updates path weights while following the conceptual map.

  In step S30, the path weight calculation unit 14 stores the concept to be calculated (referred to as concept A) in an array of starting point concepts. That is, the concept to be calculated is the initial value of the starting point concept. Also, the path weight calculation unit 14 initializes the value of D, which is a variable for counting the distance to be traced, to zero. This variable D stores the value of the path length (the number of links included in the path) at that time when going from concept to concept in the concept map. The concept processing device 1 determines the termination condition of the concept map search based on the relationship between the value of the variable D and the value of MAX_Distance, which is the parameter described above.

  Next, in step S40, the path weight calculation unit 14 sets a concept selected from the array of starting point concepts as a starting point concept (referred to as a concept C). The path weight calculation unit 14 writes the selected concept in the area storing the concept of the starting point in the memory. Then, the path weight calculation unit 14 writes all the concepts of the destination connected by the direct link from the concept of the starting point (concept C) in the array of the adjacent concepts. At this time, the path weight calculation unit 14 may obtain the concept connected to the concept C by referring to the concept map storage unit 20, or obtain the concept connected to the concept C by referring to the data of FIG. It is good.

Next, in step S50, the path weight calculation unit 14 calculates the path weight from concept A to concept C via concept A to concept C for each concept (referred to as concept B) in the array of adjacent concepts. Then, if the weight calculated for the path is larger than the weight of the path from concept A to concept B already calculated (the value of the weight calculated as MAX_Weight (A, B)), then MAX_Weight ( A and B) are updated, and the information of the path is recorded as the value of MAX_Path (A, B). The path is an extension of the path already recorded in MAX_Path (A, C) to concept B. In other words, the path is obtained by connecting the concept B to the path already recorded in MAX_Path (A, C).
If there is no calculated value as MAX_Weight (A, B), the value calculated this time is stored as the path weight value MAX_Weight (A, B). Also, the path corresponding to the weight is stored as MAX_Path (A, B).

  If the concept (concept B) extracted from the array of adjacent concepts in the process of step S50 above is the concept to be calculated (concept A) itself, the path weight calculation unit 14 determines the weight of the path. Do not calculate or record the path as MAX_Path.

  Next, in step S60, the number of paths from concept A to concept B is updated for each of the concepts B to be processed in step S50. That is, the value of NUM_Path (A, C) is added to the value of NUM_Path (A, B) up to that point, and the result of the addition is taken as the new value of NUM_Path (A, B). This corresponds to the new search from the concept A to the concept B via the concept C.

  Next, in step S70, the path weight calculation unit 14 determines whether or not an uncalculated concept remains in the array of the starting point concept. If an uncomputed concept remains in the array of starting point concepts (step S70: YES), the path weight calculation unit 14 returns to step S40 to calculate the next starting point concept. If no uncalculated concept remains in the arrangement of the starting point concept (step S70: NO), the path weight calculation unit 14 proceeds to the next step S80.

  Moving to FIG. 9, in step S80, the path weight calculation unit 14 increments the value of D, which is a variable for counting the distance.

  Then, in step S90, the path weight calculation unit 14 determines whether the value of the variable D has reached the parameter value MAX_Distance. Specifically, it is determined whether the inequality D <MAX_Distance is satisfied. When D <MAX_Distance, that is, when the distance of the path being searched has not yet reached the set maximum value MAX_Distance (step S90: YES), the process proceeds to step S100, and the process by the path weight calculation unit 14 is performed. to continue. That is, in this case, the path weight calculation unit 14 continues the process of searching for a path for the new value of D incremented in step S80 and calculating the path weight. On the other hand, when D <MAX_Distance is not satisfied, that is, when the distance of the path being searched reaches the set maximum value MAX_Distance (step S90: NO), the processing by the degree-of-association calculation unit 16 in step S110 is performed.

  Then, when the process proceeds to step S100, in the same step, the path weight calculation unit 14 writes the concepts (concept B) processed in steps S50 and S60 in the arrangement of the starting concept (see FIG. 6). Then, control is returned to step S40 (FIG. 8), and the process is repeated.

Then, when the process proceeds to step S110, in the same step, the degree-of-relevance calculating unit 16 calculates the degree of association between concepts based on the path weights calculated in the processing up to this point and the number of paths. Specifically, the degree-of-association calculation unit 16 shows the following about a concept (concept B) in which two or more routes are reached within the range set in the parameter MAX_Distance from the original concept to be calculated (concept A). Calculate the degree of association by a formula. That is
(1) The degree of association from concept A to concept B is obtained by multiplying the value of MAX_Weight from concept A to concept B by the value of the number NUM_Path of paths from concept A to concept B. That is, the degree of association from concept A to concept B is MAX_Weight (A, B) × NUM_Path (A, B).
(2) However, the degree of association calculated in the above (1) is normalized as follows. That is, the degree of association from concept A to concept B after normalization is the sum of the degree of association from concept A to concept B calculated by the equation (1) above, and the sum of the degree of association from concept A to each concept It shall be divided. That is, the degree of association after this normalization is {MAX_Weight (A, B) × NUM_Path (A, B)} / (sum of the degree of association from concept A to each concept on the concept map). By performing normalization in this manner, it is possible to calculate a value independent of the size of the concept map.

  In the process of step S110 described above, the degree of association is calculated by limiting it to a concept (concept B) in which there are two or more routes that reach from the original concept (concept A) of the original calculation object within the range set in the parameter MAX_Distance. The reasons for this are as follows. That is, the concept B, in which there is only one route from the concept A to the concept B, can be regarded as having a sufficiently low degree of association with the concept A and not related.

  Next, in step S120, the concept processing device 1 determines whether there is a concept to be calculated and a concept that has not been calculated. If there is an uncalculated concept (step S120: YES), the process returns to step S30 (FIG. 8) in order to shift to processing for the next calculation target concept. If there is an uncalculated concept (step S120: NO), the processing of the entire flowchart ends.

[Example of processing]
Next, an example of the process will be described. In the following description, reference is made to FIGS. 11, 12, 13 and 14. These drawings are schematic diagrams showing the state of the memory (work storage unit 22) in the process of the processing example described below. FIG. 11 shows the transition of values held by the storage area described in FIG. 12, 13 and 14 show the transition of values held by the storage area described in FIG. 7.

  The conceptual map targeted by the process of this example is shown in FIG. The example of the conceptual map shown in the figure includes seven concepts of "illness", "hypertension", "diabetes", "adult disease", "cold", "flu", and "fever". Each of these concepts corresponds to a node in the graph of the concept map. And, a link connecting nodes corresponds to a relation between concepts. In the graph illustrated in the same figure, there are links of disease-hypertension, disease-adult disease, disease-diabetes, disease-cold, disease-flu, adult disease-hypertension, adult disease-diabetes, cold-fever, influenza-fever include.

A processing example will be described below according to the processing procedure. The step numbers referred to in the following description of the process example are the step numbers in the procedure shown in the flowcharts of FIGS. 8 and 9.
Step S10: First, the parameter setting unit 18 of the conceptual processing device 1 sets the value of the parameter MAX_Distance to 2 in the present processing example.

  Step S20: Next, for each concept included in the concept map, the concept weight calculation unit 12 creates a list in which the directly connected concepts are listed, and calculates weights.

  FIG. 10 is a schematic view showing an example of a table for holding information on each concept created in step S20. As shown, this table has a tabular structure, and each row corresponds to a concept (node) included in the concept map. Each digit corresponds to a data item (attribute), and a node ID, a concept (node), a connecting concept (node), and a weight each exist.

The node ID is identification information conveniently assigned to each concept. In the illustrated example, numerical values 1 to 7 are assigned as node IDs.
A concept (node) is a word that expresses that concept. In this example, words such as "illness" and "adult disease" are stored in this field.
A connected concept (node) is a list of previous concepts (nodes) connected by a direct link from the concept. For example, there are five nodes linked by a direct link from the node "disease", which are "hypertension", "adult disease", "diabetes", "cold" and "flu". The connected concept column for the concept "illness" stores a list in which these five linked destination concepts are listed.
The weight stores the value of the weight for the concept (node). Since the total number of concepts appearing in this conceptual map is 7, for example, the weight for the concept "disease" (the number of concepts directly connected to 5) is log (7 / n) according to the definition log (N / n) described above. 5), that is, 0.15 (rounded to the third decimal place). Also, for example, the weight for the concept "hypertension" (the number of directly connected concepts is 5) is also log (7/2) or 0.54. The weights for other concepts are similar.

[D = 0, the concept of the origin is "illness"]
Step S30: Next, the path weight calculation unit 14 selects one concept to be calculated, and stores the concept in the area of the concept to be calculated in the memory. Also, the concept processing device 1 initializes the value of the number of concepts followed as the distance from the starting point. That is, the value of D = 0 is written in the area of the number of concepts followed in the memory. Also, the concept processing device 1 stores only the concept to be calculated selected here as an element of the array of the starting point concept in the memory.
Step S40: Next, the path weight calculation unit 14 sets the concept selected from the array of the starting point concept in the memory as the starting point concept. Then, the concept processing device 1 writes the selected concept in the area of the concept of the starting point in the memory. At present, the only element that is written to the array of origin concepts is “disease,” and therefore the concept processing device 1 writes “disease” to the area of the origin concept in the memory. Furthermore, the concept processing device 1 writes all the concepts directly connected from the concept of the origin to the array of adjacent concepts. Specifically, the concept directly connected from "disease" is "hypertension", "adult disease", "diabetes", "cold" and "influenza", so the concept processing device 1 has five of them. Write the concept of to the area of the array of adjacent concepts.
The state in the memory at the end of the above-described steps S30 and S40 is as shown in the row a of FIG.

  Next, for each concept stored in the array of adjacent concepts, the path weight calculation unit 14 sequentially performs the processing of step S50 and step S60 as described below. That is, the concept processing device 1 stores each of the concepts stored in the array of starting point concepts in FIG. 11 as the starting point concept, in the calculation target concept (in the present situation, "illness") and the array of adjacent concepts. Calculate the degree of association with the concept that is the element being The degree of association is calculated as the harmonic mean of the weight of each concept (including the concepts at both ends of the path) on the path currently being followed. Also, if the degree of association calculated here is larger than the degree of association already calculated, the information (table shown in FIG. 7) stored in the work storage unit 22 is updated, but in the present situation, the association is Since it is the first time to calculate the degree, update the path weight (MAX_Weight for each concept) for each concept. In addition, when updating the weight of this path, the path when the degree of association is calculated is also recorded (MAX_Path for each concept is updated).

(1) First, “hypertension” is taken out as concept B from the arrangement of adjacent concepts.
Step S50: The degree of association between the concept “disease” to be calculated and the concept (concept B) “hypertension” which is an element stored in the array of the adjacent concept is the weight of “disease” 0.15 (Refer to FIG. 10. The same applies to the following.) It is obtained as a harmonic mean of 0.54 which is a weight of “hypertension”. And the value is 0.23 (rounded off to the third decimal place and the same applies to the following). Also, the path corresponding to this degree of association is the path of illness-hypertension. That is, for the concept "hypertension", MAX_Weight = 0.23 and MAX_Path = "disease, high blood pressure".
Step S60: Here, the number of paths (NUM_Path) is updated. It is special only when D = 0, and here, it is calculated as NUM_Path (A, C) = NUM_Path (illness, illness) = 1. That is, for the concept "hypertension", the value of NUM_Path is updated to one.

(2) "adult disease" is taken out next as concept B from the arrangement of adjacent concepts.
Step S50: The degree of association between the concept “disease” to be calculated and the “adult disease” which is the concept B selected from the array of the adjacent concepts is 0.15 which is the weight of “disease”, “adult disease” It is obtained as a harmonic mean with 0.37, which is the weight of And the value is 0.21. Also, the path corresponding to this degree of association is the path of disease-adult disease. That is, for the concept "adult disease", MAX_Weight = 0.21 and MAX_Path = "disease, adult disease".
Step S60: As in the case of “hypertension” above, it is special only in the case of D = 0, and it is calculated as NUM_Path (A, C) = NUM_Path (illness, illness) = 1. That is, the value of NUM_Path is updated to 1 also for "adult disease".

(3) From the arrangement of adjacent concepts, "diabetes" is taken next as concept B.
Step S50: The degree of association between the concept “disease” to be calculated and “diabetes”, which is the concept B selected from the array of adjacent concepts, has a weight of “disease” of 0.15, which is “diabetes” It is obtained as a harmonic mean with a weight of 0.54. And the value is 0.23. Also, the path corresponding to this degree of association is the path of disease-diabetes. That is, for the concept "diabetes", MAX_Weight = 0.23 and MAX_Path = "disease, diabetes".
Step S60: As in the case of “hypertension” above, it is special only in the case of D = 0, and it is calculated as NUM_Path (A, C) = NUM_Path (illness, illness) = 1. That is, the value of NUM_Path is updated to 1 also about "diabetes."

(4) Next, take "cold" as concept B from the array of adjacent concepts.
Step S50: The degree of association between the concept “disease” to be calculated and “cold”, which is the concept B selected from the array of adjacent concepts, is 0.15 which is the weight of “disease” and “cold” It is obtained as a harmonic mean with a weight of 0.54. And the value is 0.23. Also, the path corresponding to this degree of association is the path of sickness-cold. That is, for "cold", MAX_Weight = 0.23 and MAX_Path = "illness, cold".
Step S60: As in the case of “hypertension” above, it is special only in the case of D = 0, and it is calculated as NUM_Path (A, C) = NUM_Path (illness, illness) = 1. That is, the value of NUM_Path is updated to 1 also for "cold".

(5) "Influenza" is taken out next as concept B from the arrangement of adjacent concepts.
Step S50: The degree of association between the concept “disease” to be calculated and “influenza”, which is the concept B selected from the array of adjacent concepts, has a weight of “disease” of 0.15 and “flu”. It is obtained as a harmonic mean with a weight of 0.54. And the value is 0.23. Also, the path corresponding to this degree of association is the path of disease-influenza. That is, for "influenza", MAX_Weight = 0.23 and MAX_Path = "illness, influenza".
Step S60: As in the case of “hypertension” above, it is special only in the case of D = 0, and it is calculated as NUM_Path (A, C) = NUM_Path (illness, illness) = 1. That is, the value of NUM_Path is updated to 1 also for "flu".

As described above, the degree of association between the “illness” that is the concept to be calculated and the concept being followed (each concept that was an element stored in the array of the adjacent concept) is calculated as the stage of D = 0. The
FIG. 12 is a schematic diagram showing the state of a table (the table shown in FIG. 7) storing the calculation results for each concept at this stage. As illustrated, MAX_Weight for each of "High blood pressure" in the first line, "adult disease" in the second line, "diabetes" in the third line, "cold" in the fourth line, and "influenza" in the fifth line. And the values of MAX_Path and NUM_Path are stored.

Step S80: Next, D is incremented and D = 1.
Step S90: Since MAX_Distance = 2 is set, the condition of D <MAX_Distance is satisfied. Thus, the process proceeds to step S100.

  Step S100: The path weight calculation unit 14 stores the concepts calculated when the loop of steps S50, S60, and S70 is repeated in the array of starting point concepts. That is, as shown in b of FIG. 11, the elements of “hypertension”, “adult disease”, “diabetes”, “cold”, and “flu” are stored in the arrangement of the starting point concept. Next, the process returns to step S40.

  Step S40: The path weight calculation unit 14 first selects the concept "hypertension" from the arrangement of starting point concepts, and sets this "hypertension" as a starting point concept (concept C). Then, all the concepts leading from this concept C (concepts adjacent to the concept C) are stored in the array of adjacent concepts. Specifically, the elements of "adult disease" and "disease" are stored in the array of adjacent concepts. This time is the state shown in b of FIG.

The processes in steps S50 and S60 are performed for "adult disease" in the array of the adjacent concept.
Step S50: The path weight calculator 14 calculates the weights for the path "illness-hypertension-adult disease" that has been traced. It is the harmonic mean of 0.15 (the weight of the concept "disease"), 0.54 (the weight of the concept "hypertension") and 0.37 (the weight of the concept "adult disease"), the value of which is 0.27 is there.
Since the value of 0.27 is larger (heavy) than the weight of 0.21 (see the second line of FIG. 12) which is the weight from "illness" to "adult disease" calculated so far, MAX_Weight ( Update A, B). In addition, MAX_Path is updated with a path corresponding to the new weight of 0.27. That is, MAX_Weight = 0.27, and MAX_Path = “disease, high blood pressure, adult disease”.
FIG. 13 is a schematic view showing the state of the table (the table shown in FIG. 7) at this time. As illustrated, MAX_Weight and MAX_Path have been updated for the concept "adult disease".

Step S60: The path weight calculation unit 14 updates the value (number of paths) of NUM_Path to NUM_Path = 2. That is, NUM_Path (disease to adult disease) at D = 1 is the sum of NUM_Path (disease to adult disease) at D = 0 and NUM_Path (disease to hypertension) at D = 0. That is, NUM_Path (A, C) at D = 0 is added to NUM_Path (A, B) at D = 0.
FIG. 14 is a schematic view showing the state of the table (the table shown in FIG. 7) at this time. As illustrated, NUM_Path is also updated for the concept "adult disease".

  In addition, since the concept "disease" is the concept itself to be calculated, calculation is not performed for the path "disease-hypertension-disease".

  Hereinafter, although the detailed description is omitted, the calculation procedure of steps S50 and S60 is repeated for the concept stored in the array of the starting point concept and the calculation of the weight is not yet completed. That is, the calculation procedure of steps S50 and S60 is repeated for each concept of “adult disease”, “diabetes”, “cold”, and “flu”. In this process, MAX_Path and NUM_Path at the time of D = 1 are obtained from the concept "illness" to each of the above concepts.

Step S80: Next, D is incremented and D = 2.
Step S90: Since MAX_Distance = 2 is set, the condition of D <MAX_Distance is not satisfied. Thus, the process proceeds to step S110.

Step S110: The degree-of-association calculation unit 16 calculates the degree of association from the concept “illness” to each concept by MAX_Weight × NUM_Path, and normalizes the value. However, this calculation is limited to the concept having two or more routes arriving from the concept "illness".
Then, the concept processing device 1 outputs the degree of association (normalized) between the concept “illness” and each concept calculated by the above-mentioned multiplication MAX_Weight × NUM_Path.
This completes the processing of the concept "illness" to be calculated.
The degree-of-association calculation unit 16 writes the calculated degree of association information in the degree-of-association storage unit 30.

  Step S120: It is determined whether or not there is an uncalculated concept to be calculated (that is, a concept to be calculated other than "ill" described above). If there is, the process proceeds to step S30 and continues. If not, the processing of the entire flowchart shown in FIGS. 8 and 9 is ended.

  The degree of association calculated by the method described above is directed. In other words, the value of the degree of association to be calculated depends on the direction from concept to concept. For example, the value of the degree of association from the concept "disease" to the concept "cold" is different from the value of the degree of association from the concept "cold" to the concept "disease".

As described above, according to the concept processing device 1 of the first embodiment, the degree of association from one concept to another can be calculated from data of a concept map. At this time, the concept processing apparatus can calculate the above-mentioned degree of association only from the data of the concept map, without requiring data such as language resources consisting of actual sentence examples.
Further, according to the concept processing device 1 of the first embodiment, the above-mentioned degree of association can be calculated with a practically appropriate limited amount of calculation. One of the factors is that the upper limit (MAX_Distance) of the distance when searching the conceptual map is provided.
Further, the degree of association between the concepts calculated by the concept processing device 1 of the first embodiment is calculated as a good value when applied to a real concept map. This is because the processing procedure and calculation method for calculation were specially devised.

  Further, according to the technology described in the first embodiment, with respect to a broad concept (for example, the concept “Japanese sports” and the like), the number of adjacent concepts is large, so that the weight of the concept becomes small. Therefore, the value of the degree of association in which such a broad concept is involved is reduced. Also, for concepts that are too narrow (for example, “NHK Sunday Sports” that represent a particular program by a particular broadcaster), the total number of paths that contain such a concept is reduced, so The value is also smaller. That is, the two extreme concepts have the effect of reducing the degree of association. That is, a low degree of association can be calculated for noise. That is, it is possible to perform processing such as information search while reducing the influence of noise in the concept.

Second Embodiment
Next, a second embodiment of the present invention will be described. In addition, description may be abbreviate | omitted about the matter which is common in the above-mentioned embodiment, and it demonstrates centering on the matter peculiar to this embodiment.
FIG. 15 is a block diagram showing a schematic functional configuration of the concept processing device according to the present embodiment. As illustrated, the concept processing device 2 includes a concept weight calculation unit 112, a path weight calculation unit 114, a degree of association calculation unit 116, a parameter setting unit 18, a concept map storage unit 20, and a work storage unit 22. , And an association degree storage unit 30; and a similarity degree acquisition unit 140.
The feature of this embodiment is that the method of calculating weights and the method of calculating relevance are different from those of the first embodiment in that the similarity acquiring unit 140 is provided.

  In the concept processing device 2 according to the present embodiment, the similarity acquisition unit 140 acquires similarity data representing the similarity between concepts from the outside. Then, the path weight calculation unit 114 calculates the path weight using the similarity between concepts by referring to the similarity data acquired by the similarity acquisition unit 140. That is, when obtaining the weight of the path that follows the concept map stored in the concept map storage unit 20, the path weight calculation unit 114 uses the value of the degree of similarity given from the outside as an initial value. That is, for the path, the path weight calculation unit 114 acquires, from the similarity data, the similarity regarding the pair of concepts connected by the links included on the path, and the path weight according to the acquired similarity as well. Calculate the value of

  FIG. 16 is a schematic view showing the configuration of similarity data acquired by the similarity acquisition unit 140. As shown in FIG. As shown, the similarity data is tabular data, and has items of concept 1, concept 2, and similarity. Concept 1 and Concept 2 correspond to the concepts contained in the concept map, respectively. The data item “similarity” holds the value of similarity between concept 1 and concept 2. In addition, as long as the similarity data includes the value of similarity between concepts as information, the format of the data may be different from the illustrated data.

  The similarity value is externally supplied, and the similarity between concepts can be given by using various methods. An example of similarity between concepts is context similarity. The method itself for determining the degree of context similarity is an existing technology. In order to determine the degree of context similarity, language resources consisting of a large amount of sentences are acquired, syntactic analysis of each sentence contained in the language resources is performed, and a statistical analysis is performed on the context in which the words appear. Here, the words correspond to concepts. In addition, the context referred to here means that, in a sentence included in a language resource, a statistical feature of a concept (word) around a concept (word) appears when a concept (word) appears, or a concept (word) Positioning in the syntax when The statistical features of surrounding concepts (words) are represented by a word vector based on surrounding concepts (words), as an example. The word vector is a multi-dimensional (eg, tens of thousands to hundreds of thousands of dimensions) vector in which each element corresponds to a word, and each element indicates whether or not the word appears in the context. Also, concepts (words) that tend to appear easily in similar contexts have higher degree of context similarity as their tendency is stronger. This is based on the premise that a concept (word) is characterized by the context, such as its surrounding concepts or syntax.

The method of determining the weight according to the present embodiment is as follows.
The concept weight calculation unit 112 calculates the weight of a concept according to the following equation.
(Weight of concept) = log (N / n)
Here, n is the number of links directly leading from the concept to another concept. Also, N is the total number of concepts included in the concept map.

Then, the path weight calculation unit 114 calculates the path weight between the adjacent concepts A and B according to any of the following equations (A1), (A2), and (A3).
Equation (A1): (weight of path between adjacent concepts) = sim (A, B) × (weight of concept A + weight of concept B)
Equation (A2): (weight of path between adjacent concepts) = sim (A, B) × (weight of concept A × weight of concept B)
Formula (A3): (weight of path between adjacent concepts) = sim (A, B) × min (weight of concept A, weight of concept B)
The path weight calculation unit 114 uses any one of the above equations (A1), (A2), and (A3), but uses any one of these equations consistently. That is, when calculating the degree of association between concepts, there is no need to change the formula for calculating the path weight between concepts on the way.

Here, sim (A, B) is the similarity between concept A and concept B, and the similarity data acquired by the similarity acquisition unit 140 can refer to the value.
Also, min (weight of concept A, weight of concept B) is a minimum value function that returns the minimum value among the arguments applied to each weight. That is, min (weight of concept A, weight of concept B) returns the smaller value of the weight of concept A or the weight of concept B.

  When the path weight calculation unit 114 calculates the degree of association from one concept (A) to another concept (B), if the two concepts are adjacent to each other, the above equations (A1), (A2) Let the value obtained by either () or (A3) be the weight as it is.

Also, when calculating weights from a certain concept (A) to another concept (B) that is not directly connected (not adjacent), the path weight calculation unit 114 performs the following calculation. That is, the harmonic mean (B1) or the geometric mean (B2) of the weights (one of formulas (A1), (A2), and (A3)) for each link included in the path connecting the two concepts is calculated. .
For example, the weight between concepts of distance 2 is calculated by either of the following equations (B1) and (B2).

Equation (B1): (weight from concept A to B) = (2 × weight from concept A to C × weight from concept C to B) / (weight from concept A to C + concept C to B) weight)
Equation (B2): (weight from concept A to B) = SQRT (weight from concept A to C × weight from concept C to B)
The path weight calculation unit 114 uses one of the above equations (B1) and (B2), but uses the selected equation among them consistently. That is, when calculating the degree of association between concepts, the formula for calculating this weight (either harmonic mean or geometric mean) is not changed on the way.

Here, concepts A to C are connected by a direct link, and concepts C to B are connected by a direct link. That is, the distance between the concepts A and B is 2, and the concept C is interposed therebetween.
Also, SQRT () in the above equation (B2) is a function that returns the square root of its argument.
Even when the distance from concept A to concept B is greater than 2, path weight calculation section 114 extends path (B1) or (B2) above to calculate the path weight by harmonic mean or geometric mean. Do.

The degree-of-association calculation unit 116 calculates the degree of association from concept A to B by any of the following methods (C1) and (C2).
(C1) The path weight from the concept A to B calculated by the path weight calculation unit 114 is output as it is as the degree of association from the concept A to B (written in the degree of association storage unit 30). When there are a plurality of paths from concept A to B, a representative path is selected by any of the following methods.
(C1-1) The shortest path is adopted.
(C1-2) The path with the largest weight (path weight) is adopted.

(C2) The largest weight of path weights from concept A to B (path weight) is multiplied by the number of paths from concept A to B, and the result is the degree of relevance from concept A to B And (output to the association degree storage unit 30). When the method (C2) is adopted, the processing procedure (procedure for searching a concept map) described in the flowchart in the first embodiment can also be used.
The degree-of-association calculation unit 116 uses one of the methods (C1) and (C2) above (in the case of (C1), one of the methods (C1-1) and (C1-2). Use the selected method among them consistently. That is, when calculating the degree of association between concepts, the method of obtaining the degree of association is not changed on the way.

  According to the concept processing device 2 of the second embodiment, using data on the degree of similarity between concepts acquired from the outside, and depending on the weight of the concept, the weight of the path, or both, The degree of association can be calculated. When the data of the degree of similarity is already provided from the outside, it is possible to obtain the degree of association that well reflects the data.

Third Embodiment
Next, a third embodiment of the present invention will be described. In addition, description may be abbreviate | omitted about the matter which is common in the above-mentioned embodiment, and it demonstrates centering on the matter peculiar to this embodiment.
A block diagram showing a functional configuration of the conceptual processing device according to the present embodiment is the same as the block diagram shown as FIG.
The concept processing device of the first embodiment multiplies the maximum value of the weight of paths between concepts (the weight of the path having the largest weight among a plurality of paths) by the number of paths between the concepts. Calculated the degree of association between the concepts. On the other hand, in the present embodiment, when there are a plurality of paths from concept A to concept B, the degree of association is obtained based on the sum of the weights of the respective paths.

Specifically, it is as follows.
The concept weight calculation unit 12 of this embodiment calculates the weight of each concept by the same method as that of the first embodiment.
Next, the path weight calculation unit 14 calculates path weights for all paths between the concept A and the concept B. The weight itself of one path is the harmonic mean of the weights of all the concepts (including the concepts at both ends) included in the path, as in the first embodiment. Note that the path between the concept A and the concept B, which is an object of calculation here, is a path that does not pass the same concept twice or more on the way. That is, a part of the path does not include a closed path or a backtracking in the middle of the path.
When the size of the concept map is finite (the number of concepts included in the concept map is finite), the number of paths between concept A and concept B is finite.
In addition, regarding the path between the concept A and the concept B to be calculated here, the length may be restricted. That is, at this time, paths which are not reached by the number of links of the value of the parameter MAX_Distance set by the parameter setting unit 18 are excluded from the calculation.

  Then, the degree-of-association calculation unit 16 obtains the sum of the weights of the paths between the concept A and the concept B, and outputs the sum as the degree of association from the concept A to the concept B. Specifically, the degree-of-association calculation unit 16 writes the calculated degree of association in the degree-of-association storage unit 30.

  According to the third embodiment, it is possible to obtain the degree of association reflecting the weight of each path, instead of representing the path between concept A and concept B by the path with the largest weight. In addition, when the size of the conceptual map is large, a large amount of calculation may be required for the process of searching for a long distance. By introducing parallel processing as needed, it is also possible to process a large amount of computation in a shorter time.

  Note that some or all of the functions of the conceptual processing device in each of the above-described embodiments may be realized by a computer and a program. In that case, a program for realizing each function may be recorded in a computer readable recording medium, and the program recorded in the recording medium may be read and executed by a computer system. Here, the “computer system” includes an OS and hardware such as peripheral devices. The term "computer-readable recording medium" refers to a storage medium such as a flexible disk, a magneto-optical disk, a ROM, a portable medium such as a ROM or a CD-ROM, or a hard disk incorporated in a computer system. Furthermore, "computer-readable recording medium" holds a program dynamically for a short time, like a communication line in the case of transmitting a program via a network such as the Internet or a communication line such as a telephone line. It may also include one that holds the program for a certain period of time, such as volatile memory in the computer system that becomes the server or client in that case. The program may be for realizing a part of the functions described above, or may be realized in combination with the program already recorded in the computer system.

  As mentioned above, although several embodiments of this invention were explained in full detail with reference to drawings, a concrete composition is not restricted to this embodiment, Design etc. of the range which does not deviate from the gist of this invention are also included. .

  The degree of association between concept "disease" and other concepts was calculated by the method described in the first embodiment, using a real concept map (different from the simplified map illustrated in the drawing). . As a result, the top ten concepts in the value of relevance are as follows. Here, the number to the fifth decimal place is calculated.

1st degree of relevance: 0.00358 concept: body 2nd degree of relevance: 0.00325 concept: inflammation 3rd degree of relevance: 0.00314 concept: medicine 4th degree of relevance: 0.00309 conception: cold 5th degree of relevance: 0 .00307 Concept: Body 6th Relevance: 0.00300 Concept: Infection 7th Relevance: 0.00296 Concept: Cancer 8th Relevance: 0.00278 Concept: Infection 9th Relevance: 0.00267 Concept: Allergy 10th relevance: 0.00258 Concept: cancer

  The present invention is applicable to various types of information processing using the degree of association between concepts. As an example, the present invention can be used in fields such as information search and program recommendation for recommending content such as a television program to a user.

1, 2 concept processing unit 12, 112 concept weight calculation unit 14, 114 path weight calculation unit 16, 116 relevance calculation unit 18 parameter setting unit 20 concept map storage unit 22 work storage unit 30 relevance storage unit 140 similarity acquisition unit

Claims (6)

Read conceptual map data that indicates the presence or absence of relationships between concepts, and calculate the value of the weight of the concept so that the smaller the number of links that represent the relationship between a concept and another concept, the larger the value of the concept. Concept weight calculation unit
In the concept map data, for a path formed by connecting a certain concept and another concept by the link, the value of the weight of the path according to the value of the weight of the concept for the concept included on the path A path weight calculation unit that calculates
A degree-of-relevance calculating unit that calculates the degree of association between the concept and another concept according to the value of the weight of the path for a path connecting a certain concept to another concept;
A concept processing apparatus comprising: The path weight calculation unit calculates the value of the weight of the path by taking a harmonic average or a geometric average of the values of the weight of the concept for each concept included on the path.
The concept processing device according to claim 1, characterized in that: The concept weight calculation unit calculates the value of the weight of the concept by calculating the logarithm of the reciprocal of the number of links between the concept and the other concept.
The concept processing apparatus according to any one of claims 1 or 2, characterized in that: When there are a plurality of paths formed by connecting the certain concept and the other concept by the link, the degree-of-association calculation unit calculates the value of the weight of the plurality of paths as the number of the plurality of paths. Calculating the degree of association between the concept and the other concept by multiplying the value of the weight of the largest one of the paths, or calculating the sum of the weights of the plurality of paths as the degree of association.
The concept processing device according to any one of claims 1 to 3, characterized in that: And a similarity acquisition unit for acquiring similarity data representing similarity between concepts;
The path weight calculation unit acquires, from the similarity data, the similarity of a pair of concepts connected by links included on the path, and the value of the weight of the path according to the acquired similarity. Calculate
The concept processing device according to any one of claims 1 to 4, characterized in that:   A program for causing a computer to function as the concept processing device according to any one of claims 1 to 5.

Images