JP5153601B2 - Search result classification apparatus, method and program - Google Patents

Description

  The present invention relates to a search result classification apparatus, method, and program for supporting search.

Conventionally, a data search system that outputs search result data as a search result, a relationship between data obtained as a search result, common feature information, and the like has been disclosed (for example, Patent Document 1).
JP 2008-146219 A

The data search system of Patent Document 1 sets a node indicating each of a plurality of search results obtained based on a database search, a link connecting nodes having common feature information, and common feature information as a link label Output information is generated by executing the generated display information generation process. Therefore, the data retrieval system of Patent Document 1, since those presented in Yu chromatography The output a lot of information including the relationship between the nodes is an output side, for example, the terminal, a large amount of information A display that can output is required. Therefore, the data search system of Patent Document 1 is particularly unsuitable for output from a mobile terminal represented by a mobile phone or the like.

  Further, the data search system of Patent Document 1 generates and stores an information database including the relationship between the links described above in advance, and extracts a search result from the information database based on the degree of coincidence. It only outputs the information already stored in the information database, which is a static database.

  The present invention provides a search result classification apparatus, method, and program capable of presenting search result information desired by a user by outputting a small amount of information particularly suitable for a mobile terminal using a dynamic database. Objective.

  The present inventor has found a mechanism for creating the minimum necessary category classification using the dynamic classification tree and presenting it to the user in an interactive manner, and has completed the present invention. Specifically, the present invention provides the following.

(1) A interactive search results classifier that outputs to the user terminal search results for input from the user terminal connected through a communication line, corresponding to the classification result of the classifying pre Symbol search results into categories the response is a rule has a search result storage unit that stores the response rule in the tree structure representing the category node, a dialogue response unit that receives an input of search criteria from the previous SL user terminal, before Symbol a search execution unit that executes a search process based on an input from the dialogue response unit, and a front Symbol search category the processed result by the execution unit classifies the category classification unit, the processing object which is classified by the prior SL category classification unit In response to the presence of a result category in the response rule, the processing target result and the search result are subordinate to the node representing the category of the response rule. Create additional child nodes, the search results and the response rule updating unit stored in the storage unit updates the response rule comprises, before Symbol dialogue response unit, wherein by referring to the response rule A search result corresponding to the search condition is output to the user terminal, the search condition based on the output search result is received, and the search corresponding to the search condition based on the search result is referred to with reference to the updated response rule the results are repeated a process of outputting to the user terminal, the search results classifier.

  According to this configuration of the present invention, the processing target result obtained by executing the interactive search process on the response rule corresponding to the classification result obtained by classifying the search result into categories and representing the tree-structured response rule representing the category as a node. Can be updated and reflected. Therefore, it is possible to perform a search process for a result that has already been searched and classified using a dynamic database in which the processing target result is updated in the response rule. In addition, the minimum necessary category classification can be created by a dynamic classification tree in which search results are classified into categories, and processing target results are created in categories that match the classification.

(2) The search result accumulation unit changes the other upper node except for the one upper node among a plurality of upper nodes having the same lower node to a degenerated node that aggregates the same lower nodes. stored and, before Symbol response rule updating unit, in response to the previous SL category of the processed results are classified by the category classification unit is not present in the response rule, includes a category of the processed result, in response to the deepest node search unit for searching the deepest nodes located at the bottom of the response rule, before Symbol deepest node the deepest node searched by the searching unit does not have a node in the lower, the deepest node There in response to the the degeneracy determination unit determines whether a degenerate node was determined to be the degenerate node by prior Symbol degeneracy determination unit, to the deepest node, the reduced Has a degeneracy canceller for generating the lower node corresponding to retreat node, and the lower of the deepest node searched by pre Symbol deepest node searching unit, the child to add the processed result as the search result create a node, to update the response rule stored in the search result accumulation unit, the search result classification apparatus according to (1).

According to such a configuration of the present invention, the response rule is changed from a plurality of upper nodes having the same lower node other than the upper node to a degenerate node in which the same lower nodes are aggregated. To do. Thus, by aggregating upper node having a same lower node, it is possible to reduce the capacity of the storage area of the search result accumulation unit that stores a response rule. Further, it is possible to reduce processing time for an unnecessary search at the time of search. Furthermore, since only the degenerated node necessary for adding the processing target result is eliminated and the processing target result is added, the processing target result can be added with a minimum of processing.

(3) the interaction response unit has a front Symbol category receiving section that receives an input of the category, before Symbol search results with reference to the stored the response rule in the storage unit, corresponding to the category to which the category accepting unit accepts the categories and a result presentation unit that outputs the label of the lower node of the node, before Symbol category receiving unit, including the said selected by the user terminal, the label on the basis of an output by the result presentation unit for accepting an input of the pre-Symbol result presentation unit, you output label of the lower of the nodes of the node corresponding to the category that matches the category the category accepting unit accepts, (1) or (2) The search result classification device described in 1.

  According to such a configuration of the present invention, the label of the lower node of the node corresponding to the received category is output, and the label of the lower node is output by receiving the category based on the output. The categories received based on the output are included in the lower node labels. Therefore, using the response rule tree structure, it is possible to drill down from the upper level to the lower level to obtain the result, and it is possible to provide a mechanism that is not limited to the fixed search intention. In addition, it is possible to provide a mechanism that leads to search results interactively.

(4) the result presentation unit, you output the label is limited documents below a predetermined number of characters represented by the word, the search results classification apparatus according to (3).

  According to such a configuration of the present invention, since a document limited to a predetermined number of characters typified by words is output as a label of a lower node, the document can be output with a small number of words. Even a device having a small display can efficiently output search results to the user.

(5) the interaction response unit, a processing request receiving unit which receives a search processing request from the search processing request received by the pre-Symbol process request receiving unit, the search is the input information of the execution unit search instruction request A search result classification device according to any one of ( 1) to (4) , comprising: a query extraction unit for extraction.

  According to such a configuration of the present invention, since a search instruction request represented by a query is extracted from the search processing request, a query necessary for the search can be extracted from a search processing request such as a keyword that the user wants to search.

(6) An interactive computer that outputs a search result to an input from a user terminal connected via a communication line to the user terminal is a response rule corresponding to a classification result obtained by classifying the search result into a category. A search result storage unit that stores the response rule of a tree structure in which a category is represented by a node, and wherein the computer receives an input of a search condition from the user terminal ; and a search execution step of executing a search process based on an input from the dialogue response step, the computer, and a category classification step of classifying the processed result by the search execution step to the category, the computer, by the categorization step The classified category of the processing target result exists in the response rule. The response stored in the search result storage unit is created by creating a child node that adds the processing target result as the search result below the node representing the category of the response rule. A response rule update step for updating a rule, and the computer outputs a search result corresponding to the search condition to the user terminal with reference to the response rule, and accepts a search condition based on the output search result, results and the step of repeating the process of outputting to the user terminal, the including, search result classification method corresponding to the search condition based on the search result by referring to the updated response rule.

  (7) A search result classification program for causing a computer to execute the steps of the method according to (6).

  According to the present invention, it is possible to perform a search process and a result presentation process in parallel by a dialog using a dynamic database in which a processing target result is updated in a response rule.

  Hereinafter, the best mode for carrying out the present invention will be described with reference to the drawings. This is merely an example, and the technical scope of the present invention is not limited to this.

(Embodiment)
[Overall Configuration of Search System 100 and Functional Configuration of Search Result Classification Server 1]
FIG. 1 is a diagram showing the overall configuration of the search system 100 and the functional configuration of the search result classification server 1 according to the present embodiment.

  The search system 100 includes a search result classification server 1, a plurality of content servers 2, a communication network 3, and a terminal 4. As shown in FIG. 1, a search result classification server 1, a content server 2, and a terminal 4 are connected to each other via a communication network 3 typified by the Internet or the like.

  The search result classification server 1 is an interactive search support apparatus that receives a search processing request from the terminal 4 and outputs a search result for the search processing request to the terminal 4 in an interactive format. The search result classification server 1 includes a control unit 10 and a storage unit 20. The control unit 10 includes a dialogue response unit 11 and an execution unit 16. The dialogue response unit 11 includes a processing request reception unit 12, a query extraction unit 13, a result presentation unit 14, and a user instruction reception unit 15. The execution unit 16 includes a search execution unit 17, a category classification unit 18, and a response rule update unit 19. The response rule update unit 19 further includes a deepest node search unit 19A, a degeneration determination unit 19B, and a degeneration cancellation unit 19C. The storage unit 20 includes a search result storage database (hereinafter, “database” is also referred to as “DB”) 22 which is a search result storage unit.

The search result classification server 1 is not limited in the number of hardware, and may be configured by one or a plurality of hardware as necessary. Further, when the search result classification server 1 is configured by a plurality of hardware, the hardware may be connected via the communication network 3. For example, a separate server for each capacity machine you later, the transmission and reception of signals between the server, by linking each server may implement the functions of the present embodiment.

  The dialogue response unit 11 serves as an interface between the terminal 4 and the execution unit 16. Further, the dialog response unit 11 refers to the search result accumulation DB 22 and outputs the search result to the terminal 4 in an interactive manner. The execution unit 16 executes a search process based on the search query acquired from the dialogue response unit 11. Then, the execution unit 16 classifies the results of the search process into categories and updates the search result storage DB 22 by placing the search processing results in a lower category than the corresponding category in the search result storage DB 22. A category refers to a category to which items of the same nature such as “entertainment”, “celebrity”, “talent”, and “child role” belong, and is set in advance. Details will be described later.

  The search result accumulation DB 22 is a DB that classifies the search results into categories and stores search response rules corresponding to the classification results. The response rule can be expressed as a tree structure in which categories and search results are represented by nodes. Response rules can be implemented using an extended semantic network technique that extends the semantic network. Here, the semantic network is a representation of the meaning of sentences and words handled as one block in the program, associative relationships between concepts, knowledge, etc. in a network format using a known graph structure. Say. A semantic network is a correlation diagram in which individual concepts are nodes and their semantic relations are connected by arrows (edges). By using a semantic network technique, computer processing is facilitated. The extended semantic network refers to a combination of two elements in which a combination in which both the start point and the end point are empty elements is treated as a node, and a combination in which both the start point and the end point are not empty elements is treated as an edge. Therefore, the extended semantic network includes a network whose start point is a node and whose end point is an edge.

  The content server 2 is a content DB server that stores a plurality of Web pages. If the content server 2 is connected to the communication network 3, it corresponds to a DB server for any content that is arranged all over the world and stores a Web page.

Terminal 4 is a device having a push button for a user to an operation input for search, a keyboard, an input unit and a mouse, a display screen for displaying the content. For example, the terminal 4 has a small display such as a mobile phone 4a and is not suitable for handling a large amount of information, an information communication terminal such as a PDA (Personal Data Assistant) 4b, or a PC (personal computer) 4c. It may be. Using the terminal 4, the user can perform a search process.

[Hardware configuration diagram of search result classification server 1]
FIG. 2 is a diagram illustrating a hardware configuration of the search result classification server 1 according to the present embodiment. The server in which the present invention is implemented may be a standard server, and an example of the configuration is shown below.

The search result classification server 1 includes a plurality of CPUs such as a CPU (Central Processing Unit) 1 (hereinafter referred to as “CPU1 010 ” ) constituting the control unit 10 and a CPU 2 (hereinafter referred to as “CPU 1012 ”) in a multiprocessor configuration. Bus line 1005, communication I / F (I / F: interface) 1040, main memory 1050, BIOS (Basic Input Output System) 1060, display device 1022, I / O controller 1070, keyboard and mouse, etc. input device 1100 includes a hard disk 1074, optical disk drive 107 6 semiconductor memory 1078 to the parallel beauty. Incidentally, a hard disk 1074, optical disk drive 107 6及 beauty semiconductor memory 1078, collectively referred to as the storage unit 20.

  The control unit 10 is a part that controls the search result classification server 1 in an integrated manner. By appropriately reading and executing various programs stored in the hard disk 1074, the control unit 10 cooperates with the hardware described above, and relates to the present invention. Various functions are realized.

  The communication I / F 1040 is a network adapter when the search result classification server 1 transmits / receives information to / from the terminal 4 (FIG. 1) via the communication network 3 (FIG. 1). The communication I / F 1040 may include a modem, a cable modem, and an Ethernet (registered trademark) adapter.

  The BIOS 1060 records a boot program executed by the CPU 1010 when the search result classification server 1 is started, a program depending on the hardware of the search result classification server 1, and the like.

  The display device 1022 includes a display device such as a cathode ray tube display device (CRT) or a liquid crystal display device (LCD).

The I / O controller 1070 may be connected to the hard disk 1074, optical disk drive 107 6及 beauty storage unit 20 is a storage device such as a semiconductor memory 1078.

  The input device 1100 is for accepting input by an administrator of the search result classification server 1.

  The hard disk 1074 stores various programs for causing the hardware to function as the search result classification server 1, a program for executing the functions of the present invention, and the search result accumulation DB 22 described above. The search result classification server 1 can also use an external hard disk (not shown) as an external storage device.

As the optical disk drive 1076, for example, a DVD-ROM drive, a CD-ROM drive, a DVD-RAM drive, and a CD-RAM drive can be used. In this case , the optical disk 1077 corresponding to each drive is used. A program or data may be read from the optical disk 1077 by the optical disk drive 1076 and provided to the main memory 1050 or the hard disk 1074 via the I / O controller 1070.

  Note that the computer in the present invention refers to an information processing device including a storage device, a control unit, and the like, and the search result classification server 1 is configured by an information processing device including the storage unit 20, the control unit 10, and the like. This information processing apparatus is included in the concept of the computer of the present invention.

[Hardware Configuration of Content Server 2]
The content server 2 also has the same configuration as the search result classification server 1 described above. Not only the content server 2 but also the search result classification server 1 may have a part of the content DB that stores the Web page in the same manner as the content server 2.

[Hardware configuration of terminal 4]
The terminal 4 also has the same configuration as the search result classification server 1 described above. For example, the storage unit may be a drive in which an external memory can be inserted instead of the optical disc drive 1076.

[Data structure of search result storage DB 22]
Figure 3 is an example of a knowledge data structure (response rules) 30 of the search result accumulation DB22 according to the present embodiment is a view to view. The knowledge data structure 30 in FIG. 3 relates to the search for “curry”. As an example, the knowledge data structure 30 includes a category K (K1 to K6), a search result M (M1 to M11), an option O (O1, O2), and a question Q (Q1 to Q3). Category K, search result M, option O, and question Q can all be represented as nodes. A line connecting nodes is an edge. There are two types of edges, one indicating a flow (indicated by a solid line in the figure) and one indicating a selection message (indicated by a broken line in the figure). The knowledge data structure 30 is represented by a node and an edge, and has a tree structure in which the node can be traced from upper to lower.

  The category K is a category for classifying search results, and a heading (label) indicating the characteristics of the lower nodes is attached. In the example of FIG. 3, the label is a document limited to a predetermined number of characters (for example, 20 characters) such as “curry” and “Roppongi”. The search result M is a message output as a search result. Option O is to associate categories K with each other in order to determine category K. Since the option O determines the category K, the question Q corresponding to the category K is associated. The question Q is a question for determining the category K and deriving the search result M to the minimum.

[Relationship between knowledge data structure 30 and search result classification server 1]
FIG. 4 is a diagram showing the architecture 40 of the search result classification server 1 according to the present embodiment using the knowledge data structure 30.

  The search result classification server 1 includes a computer program for realizing a dialog processing function executed by the dialog response unit 11 (see FIG. 1). The computer program is stored in, for example, the hard disk 1074 in FIG. 2, and is created in the main memory 1050 when the system is activated.

  In addition, the computer program includes a kernel 41 related to operations such as basic input / output of the system, an Internet Relay Chat (Internet Relay Chat), which is a name of a framework in which the search result classification server 1 communicates with the terminal 4. Session management 42 for IRC), and channel management 43 and 44 in IRC.

  The channel management 43 and 44 may assign one channel management to one user (terminal 4), or may assign a plurality of channel managements to one user. The correspondence relationship between the user and the channel management is not particularly limited. The channel managements 43 and 44 are not limited to these, and an arbitrary number of channel managements may be provided. The number of responses and the like that can be controlled by each channel management can be designed as appropriate.

  The computer program appropriately includes a basic response (DST: Dialogue Script Tracer) 51 that can be controlled by the channel management 43, a speech control response 52, a help response 53, and the like, and a basic response that can be controlled by the channel management 44. (DST) 54, time response 55, etc. are included as appropriate. These are response modules having individual functions, and the function of each response module can be designed as appropriate.

The basic response (DST) 51, 54 includes generating a basic response to the input from the user in the dialog response, specifically, generating a typical response at the start or end of the dialog, etc. Including. The basic responses (DST) 51, 54 also include the generation of responses with constraints, such as showing the user finite choices in the dialogue response. The speech control response 52 provides a function that allows the user to arbitrarily suspend / resume a long response message. The help response 53 includes appropriately generating a response such as help information in response to an operation explanation request from the user. The time response 55 includes generation of a response based on the elapsed time of the dialog or the operating time of the system, and specifically, to send the elapsed time such as “used for 30 minutes” to the user at the end. It is provided, or includes, etc. to be prepared in order to send a "thank you" Jiko of greeting such as the like to the user in the year beginning.

  The inference engine 45 is a mechanism for deriving a search result as an answer from the knowledge data structure 30. The search thread 46 corresponds to the execution unit 16 (see FIG. 1).

  Further, the computer program can access the knowledge data structure 30 using the basic responses (DST) 51 and 54, the speech control response 52, and the like.

  The search result classification server 1 accesses the knowledge data structure 30 using the basic responses (DST) 51 and 54, and generates data for the interactive response using the data already stored in the knowledge data structure 30. Interactive processing can be performed. Further, the search result classification server 1 can update the search result and store it in the knowledge data structure 30 when the search thread 46 accesses the knowledge data structure 30.

[Interactive processing]
FIG. 5 is a flowchart of dialogue processing according to the present embodiment. The dialogue processing means that the search result classification server 1 derives a search result desired by the user of the terminal 4 by transmitting / receiving data to / from the terminal 4. This process is a process performed by the search result classification server 1 until data is received from the terminal 4 and data is transmitted to the terminal 4.

  S1: The control unit 10 (the dialogue response unit 11) accepts an input from the terminal 4.

  S2: The control unit 10 (processing request receiving unit 12) determines whether or not the received request is a search processing request. The search processing request is, for example, processing when a search is made for the first time when a declaration such as “search for the next word” is made from the terminal 4 and a keyword or the like is input on the search screen displayed on the terminal 4. A request. If the request is a search process request (S2: YES), the process request receiving unit 12 moves the process to S3. On the other hand, if the request is not a search process request (S2: NO), the process request accepting unit 12 moves the process to S6.

  S3: The control unit 10 (query extraction unit 13) extracts a search instruction request from the received search processing request. The search instruction request refers to a query for performing an Internet search with an external Web search service (not shown) connected via the communication network 3.

  S4: The control part 10 (execution part 16) performs the search process mentioned later. This search process is merely a query acceptance process, and the process of actually performing a Web search and updating the response rule is executed in parallel with the interactive process.

S5: The control unit 10 (result presentation unit 14) transmits the processing result to the terminal 4. Here, the processing result is, for example, refers to a "specified query has been registered with the web search process queue." And the like. Then, the control part 10 complete | finishes this process.

  S6: The control unit 10 (user instruction receiving unit 15) determines whether or not an instruction from the user such as “Please present the search result” is received from the terminal 4. When receiving an instruction from the user (S6: YES), the user instruction receiving unit 15 moves the process to S7. On the other hand, when the instruction from the user has not been received (S6: NO), the control unit 10 ends this process.

  S7: The control unit 10 (the dialogue response unit 11) performs a DST process to be described later. Then, the control part 10 complete | finishes this process.

[Search processing]
FIG. 6 is a flowchart of search processing according to this embodiment. Since the search process is a time-consuming process, it can be processed as a separate process or thread. Here, processing of one thread is shown.

  S11: The control unit 10 (search execution unit 17) performs an Internet search using an external Web search service using the extracted query. Specifically, the control unit 10 (search execution unit 17) executes a search API (Application Program Interface) using the extracted query.

  S12: The control unit 10 (category classification unit 18) categorizes the search result that is the execution result of the search API.

  S13: The control part 10 (response rule update part 19) performs the response rule update process mentioned later.

  S14: The control unit 10 (execution unit 16) determines whether or not the search has been executed until a predetermined number (for example, 10,000) is satisfied. The predetermined number means the maximum number of cases for which a search result is acquired. Normally, the search API does not acquire a large amount of search results at once, and therefore, iterates. When the search is performed until the predetermined number is satisfied (S14: YES), the execution unit 16 ends the process. On the other hand, when the search is not performed until the predetermined number is satisfied (S14: NO), the execution unit 16 moves the process to S11.

[Response rule update processing]
FIG. 7 is a flowchart of the response rule update process according to the present embodiment.

  S21: The response rule update unit 19 searches the search result category from the response rule of the search result accumulation DB 22.

  S22: The response rule update unit 19 determines whether or not a corresponding category is found in the response rule. If the category is found (S22: YES), the response rule update unit 19 moves the process to S23. On the other hand, when the category is not found (S22: NO), the response rule update unit 19 moves the process to S24.

  S23: The response rule update unit 19 sets the node corresponding to the found category as the node n. Thereafter, the response rule update unit 19 moves the process to S31.

  S24: The response rule update unit 19 (the deepest node search unit 19A) searches the response rule for the deepest node n whose search result category matches from the top. That is, the deepest node n is a node that includes the category of the search result to be updated and is positioned at the lowest level of the response rule.

  S25: The response rule update unit 19 counts the number of children of the deepest node n in the response rule.

  S26: The response rule update unit 19 determines whether or not the number of children of the deepest node n is zero. The child of the deepest node n refers to a node other than the degenerated node described later among the nodes below the deepest node n. If the number of children of the deepest node n is 0 (S26: YES), the response rule update unit 19 moves the process to S27. On the other hand, when the number of children of the deepest node n is not 0 (S26: NO), the response rule update unit 19 moves the process to S30.

  S27: The response rule update unit 19 (degeneration determination unit 19B) determines whether or not the lowermost of the deepest node n is a degenerate node. If the subordinate of the deepest node n is a degenerate node (S27: YES), the response rule update unit 19 moves the process to S28. On the other hand, when the subordinate of the deepest node n is not a degenerate node (S27: NO), the response rule update unit 19 moves the process to S30.

  S28: The response rule update unit 19 (degeneration canceling unit 19C) creates a child node in the deepest node n and cancels the degenerate node.

  S29: The response rule update unit 19 further searches for the category of the search result from the response rule, creates a child node for the found node, and makes it a node n. Thereafter, the response rule update unit 19 moves the process to S31.

  S30: The response rule update unit 19 creates a child node of the deepest node n and sets the child node to the node n.

  S31: The response rule update unit 19 adds the search result to the node n and updates the response rule.

  As described above, the search result classification server 1 can update and reflect the search result in the corresponding category of the response rule having the tree structure. In addition, the search result classification server 1 can classify the search results into categories, and create a minimum category classification by using a dynamic classification tree in which the search results are further created for the categories that match the classification.

Furthermore , since the search result classification server 1 eliminates only the degenerate nodes necessary for adding the search result and adds the search result, the search result can be added with a minimum of processing.

[Specific example of response rule update processing]
Here, the response rule update processing will be described based on a specific example. FIG. 8A shows the response rules stored in the search result accumulation DB 22. The search process is performed using a query “DNA microarray test”. Suppose that a search result for the query “DNA microarray assay” is to be added to the “biology” category of “natural sciences and technology”. Further, in the response rule shown in FIG. 8A, the node (# 2) of the “statistical terminology dictionary” subordinate to the node (# 1) of the category “natural science and technology” is a degenerate node. A degenerate node is a node in which when a node and its subordinate nodes are the same as one or more other nodes, other nodes than any one node are identified as normal nodes. Say.

  First, the control unit 10 searches for the category of the search result from the response rule (S21 in FIG. 7), but the corresponding category is not found because it is a degenerate node. Therefore, the control unit 10 searches for the “natural science and technology” node (# 1) as the deepest node in which the category of the search result matches from the top (S24 in FIG. 7). The “natural science and technology” node is a category node, and since only the degenerate node that is the node (# 2) of the “statistical terminology dictionary” exists at the lower level, the number of children is zero.

  FIG. 8B shows the response rule after eliminating the degenerate node. Since the node of the “statistical terminology dictionary” is a degenerate node, the control unit 10 creates a corresponding child node (# 3). Since there is a category corresponding to the search result for the query “DNA microarray test” (# 4), the control unit 10 creates its child node (# 5) and adds the search result (FIG. 7). S28, S29, S31). The search results “Statistics Dictionary” and “DNA Microarray” are common up to the category “Natural Science and Technology”, and the categories “Mathematics” and “Biology” below it have the path of classification results for the first time. These processes were different, and these processes could construct a minimal tree structure that represented their results.

[DST processing]
FIG. 9 is a flowchart of the DST process according to the present embodiment.

  S41: The dialogue response unit 11 determines whether or not a context pointer described later is defined. If the context pointer is defined (S41: YES), the dialog response unit 11 moves the process to S43. On the other hand, when the context pointer is not defined (S41: NO), the dialog response unit 11 moves the process to S42. Note that the context pointer may be stored in the storage unit 20. Further, the context pointer is undefined when the DST process is first performed among the interactive processes (FIG. 5) repeated a plurality of times.

  S42: The dialogue response unit 11 searches for a node that matches the category specified by the user's instruction from the top of the response rule. Thereafter, the dialog response unit 11 moves the process to S44.

  S43: The dialogue response unit 11 searches for a node that matches the category specified by the user from the nodes in which the context pointer is defined.

  S44: The dialog response unit 11 determines whether a matching node is found. If a matching node is found (S44: YES), the dialog response unit 11 moves the process to S45. On the other hand, when a matching node is not found (S44: NO), the dialogue response unit 11 ends this process.

  S45: The dialog response unit 11 selects one matched node. The selection when there are a plurality of matched nodes may be selected at random, or may be performed most recently. It may also be the oldest executed.

  S46: The dialogue response unit 11 determines whether or not the lower node of the selected node is the search result M. If the search result is M (S46: YES), the dialogue response unit 11 moves the process to S47. On the other hand, if it is not the search result M (S46: NO), the dialog response unit 11 moves the process to S48.

  S47: The dialogue response unit 11 (result presentation unit 14) responds with the search result M as a message. If there are a plurality of search results M, the plurality of search results M may be returned as a message. Thereafter, the dialog response unit 11 ends this process.

  S48: The dialogue response unit 11 determines whether or not the lower node of the selected node is the option O. If the option is O (S48: YES), the dialogue response unit 11 moves the process to S49. On the other hand, when it is not the option O (S48: NO), the dialog response part 11 complete | finishes this process.

  S49: The dialogue response unit 11 sets the node of option O as the context pointer.

  S50: The dialogue response unit 11 (result presentation unit 14) responds with the question Q associated with the option O and all the categories K that are child nodes of the option O. Thereafter, the dialog response unit 11 ends this process.

  As described above, the search result classification server 1 repeatedly executes the DST process to output the labels of the lower nodes of the node corresponding to the category received from the terminal 4, and sets the category to the same terminal 4 based on the output. By receiving from, the label of the lower node is output. Therefore, the search result classification server 1 can use the response rule tree structure to drill down from the upper level to the lower level and output the search result desired by the user in an interactive manner.

[Display example of mobile phone 4a]
FIG. 10 is a diagram showing a display example on the mobile phone 4a according to the present embodiment.

  First, FIG. 10A shows a search screen 60 displayed on the display of the mobile phone 4a. The search screen 60 includes a search condition input unit 61 and a search button 62. In the search condition input unit 61, the word “DNA microarray test” is input as a search keyword. Further, by selecting the search button 62, the data input to the search condition input unit 61 is transmitted to the search result classification server 1.

  As shown in FIG. 10 (2), a system response unit 63 and an option display unit 64 are displayed on the search screen 60. The system response unit 63 is an interactive question from the search result classification server 1. In this example, the system response unit 63 is a question (not shown) associated with “(Option A)” in FIG. The option display unit 64 displays the category K associated with “(Option A)”. When there are a plurality of associated categories K, all of them may be displayed on the option display unit 64. For example, the user selects “natural science and technology” from the option display section 64.

  FIG. 10 (3) shows a search screen 60 output to the mobile phone 4a after the user selects “Natural Science and Technology” (FIG. 10 (2)). The search screen 60 displays a question (not shown) associated with “(Option B)” in FIG. 8B and the category K associated with “(Option B)”. When the user selects “biology”, the search screen 60 shown in FIG. 10 (4) is displayed. Thereafter, the search result desired by the user in the interactive form between the search result classification server 1 and the mobile phone 4a. Can be output.

As mentioned above, although embodiment of this invention was described, this invention is not restricted to embodiment mentioned above. Embodiments The effects described in the present invention is merely a list of the most preferable effects resulting from the present invention, the effect of the present invention be limited to those described in the embodiments of the present invention is not.

It is a figure which shows the whole structure of the search system which concerns on this embodiment, and the function structure of a search result classification server. It is a figure which shows the hardware constitutions of the search result classification server which concerns on this embodiment. It is a figure which shows the example of the knowledge data structure (response rule) of search result storage DB which concerns on this embodiment. It is a figure which shows the architecture of the search result classification server which concerns on this embodiment using a knowledge data structure. It is a flowchart of the dialogue process which concerns on this embodiment. It is a flowchart of the search process which concerns on this embodiment. It is a flowchart of the response rule update process which concerns on this embodiment. It is a figure which shows the specific example of the response rule update process which concerns on this embodiment. It is a flowchart of the DST process which concerns on this embodiment. It is a figure which shows the example of a display with the mobile telephone which concerns on this embodiment.

DESCRIPTION OF SYMBOLS 1 Search result classification server 4 Terminal 10 Control part 11 Dialog response part 12 Process request reception part 13 Query extraction part 14 Result presentation part 15 User instruction reception part 16 Execution part 17 Search execution part 18 Category classification part 19 Response rule update part 19A Deepest Node search unit 19B Degeneration determination unit 19C Degeneration cancellation unit 20 Storage unit 22 Search result accumulation DB
30 Knowledge Data Structure 60 Search Screen 100 Search System K Category M Search Result O Choice Q Question

Claims (6)

An interactive search result classification device for outputting a search result for an input from a user terminal connected via a communication line to the user terminal,
A search result storage unit that stores the response rule of a tree structure that is a response rule corresponding to a classification result obtained by classifying the search result into a category and that represents the category as a node;
A dialog response unit that accepts input of search conditions from the user terminal;
A search execution unit that executes a search process based on an input from the dialogue response unit;
A category classification unit for classifying the processing target result by the search execution unit into the category;
When the category of the processing target result classified by the category classification unit exists in the response rule, the processing target result is added as the search result to the lower level of the node representing the category of the response rule. A response rule update unit that creates a child node to update the response rule stored in the search result storage unit;
With
The interactive response unit refers to the response rule, outputs a search result corresponding to the search condition to the user terminal, accepts the search condition based on the output search result, and refers to the updated response rule the search results corresponding to search results to the search criteria based to repeat the process of outputting to the user terminal and,
The search result accumulating unit stores the upper nodes other than one upper node among a plurality of upper nodes having the same lower node by changing the upper nodes other than the upper node to a degenerated node in which the same lower nodes are aggregated. And
The response rule update unit
In response to the fact that the category of the processing target result classified by the category classification unit does not exist in the response rule, a search is made for the deepest node that includes the category of the processing target result and is positioned at the lowest level of the response rule. The deepest node search unit;
A degeneration determination unit that determines whether or not the deepest node is the degenerate node in response to the deepest node searched by the deepest node search unit not having a lower node;
A degeneration resolving unit that creates the lower node corresponding to the degenerated node in the deepest node in response to the degeneration determining unit determining that the node is the degenerated node;
The child node that adds the processing target result as the search result is created below the deepest node searched by the deepest node search unit, and the response rule stored in the search result storage unit is updated. ,
Search result classification device. The dialogue response unit includes:
A category reception unit for receiving input of the category;
A result presentation unit that refers to the response rule stored in the search result storage unit and outputs a label of a node below the node corresponding to the category received by the category reception unit;
With
The category receiving unit receives an input of the category included in the label, selected by the user terminal based on an output from the result presentation unit,
The result presentation unit outputs a label of the node below the node corresponding to the category that matches the category received by the category reception unit,
The search result classification apparatus according to claim 1. The result presentation unit outputs the label that is a document limited to a predetermined number of characters represented by a word,
The search result classification device according to claim 2 . The dialogue response unit includes:
A process request accepting unit for accepting a search process request;
A query extraction unit that extracts a search instruction request that is input information of the search execution unit from the search processing request received by the processing request reception unit;
Comprising
The search result classification device according to any one of claims 1 to 3 . An interactive computer that outputs a search result for an input from a user terminal connected via a communication line to the user terminal is a response rule corresponding to a classification result obtained by classifying the search result into a category, and the category is a node A search result storage unit for storing the response rules of the tree structure represented by
An interactive response step in which the computer receives an input of a search condition from the user terminal;
A search execution step in which the computer executes a search process based on an input from the dialog response step;
A category classification step in which the computer classifies the processing target result of the search execution step into the category;
In response to the fact that the category of the processing target result classified by the category classification step exists in the response rule, the computer places the processing target result below the node representing the category of the response rule. A response rule update step of creating a child node to be added as a search result and updating the response rule stored in the search result storage unit;
The computer refers to the response rule, outputs a search result corresponding to the search condition to the user terminal, accepts a search condition based on the output search result, refers to the updated response rule Repeating the process of outputting a search result corresponding to a search condition based on the search result to the user terminal;
Only including,
The search result accumulating unit stores the upper nodes other than one upper node among a plurality of upper nodes having the same lower node by changing the upper nodes other than the upper node to a degenerated node in which the same lower nodes are aggregated. And
The response rule update step includes:
In response to the fact that the category of the processing target result classified by the category classification step does not exist in the response rule, a search is made for the deepest node that includes the category of the processing target result and is positioned at the lowest level of the response rule. Deepest node search step;
A degeneration determination step of determining whether or not the deepest node is the degenerate node in response to the deepest node searched in the deepest node search step having no lower node;
A degeneration cancellation step of creating, in the deepest node, the lower node corresponding to the degenerated node in response to the degeneration determining step determining that the node is the degenerated node;
The child node that adds the processing target result as the search result is created below the deepest node searched by the deepest node search step, and the response rule stored in the search result storage unit is updated. ,
Search result classification method. A search result classification program for causing a computer to execute the steps of the method according to claim 5 .

Images