JP7626440B2 - Information processing system, information processing method, and program - Google Patents

Description

The present invention relates to an information processing system, an information processing method, and a program.

Patent document 1 discloses a causal relationship extraction system that extracts causal relationships between economic events and corporate performance.

JP 2020-24689 A

In recent years, there has been a demand for more quantitative predictions of various economic phenomena using information processing technology. However, although the technology disclosed in Patent Document 1 can extract causal relationships, it cannot make quantitative predictions such as increases or decreases in parameters.

In consideration of the above circumstances, the present invention aims to provide an information processing system etc. that allows users to more easily obtain quantitative prediction results regarding various economic phenomena etc.

According to one aspect of the present invention, an information processing system is provided. This information processing system is configured to execute the following steps. In the acquisition step, input information received from a user and preset reference information are acquired. The reference information is information including causality that associates multiple causes and effects. In the extraction step, a related numerical indicator is extracted based on the acquired input information and reference information. The related numerical indicator is related to at least one item that appears in a chain of causal relationships starting from a keyword included in the input information.

According to this embodiment, users can more easily obtain quantitative prediction results regarding various economic phenomena, etc.

1 is a configuration diagram illustrating an information processing system 1 according to an embodiment of the present invention. FIG. 2 is a block diagram showing the hardware configuration of a user terminal 2. FIG. 2 is a block diagram showing a hardware configuration of the server 3. 2 is a block diagram showing functions realized by a control unit 33 and the like in the server 3. FIG. FIG. 11 is an activity diagram showing the flow of information processing relating to generation of reference information IF. FIG. 13 is an activity diagram showing the flow of information processing regarding the presentation of the causal chain 4 and the related numerical index RI. FIG. 13 is a conceptual diagram for explaining a causal chain 4. 2 is a diagram showing an example of a screen 5 of a display unit 24 visually recognized by a user. FIG. 2 is a diagram showing an example of a screen 5 of a display unit 24 visually recognized by a user. FIG. 11A to 11C are diagrams showing examples of different aspects of the screen 7 of the display unit 24 visually recognized by the user. 11A to 11C are diagrams showing examples of different aspects of the screen 7 of the display unit 24 visually recognized by the user.

The following describes embodiments of the present invention with reference to the drawings. The various features shown in the following embodiments can be combined with each other.

The program for realizing the software appearing in this embodiment may be provided as a non-transitory computer-readable recording medium, or may be provided so that it can be downloaded from an external server, or may be provided so that the program is started on an external computer and its functions are realized on a client terminal (so-called cloud computing).

In this embodiment, a "unit" may also include, for example, a combination of hardware resources implemented by a circuit in the broad sense and software information processing that can be specifically realized by these hardware resources. In addition, this embodiment handles various types of information, which may be represented, for example, by physical values of signal values representing voltage and current, high and low signal values as a binary bit collection consisting of 0 or 1, or quantum superposition (so-called quantum bits), and communication and calculations may be performed on a circuit in the broad sense.

In the broad sense, a circuit is a circuit realized by at least appropriately combining a circuit, circuitry, a processor, and memory. In other words, it includes application specific integrated circuits (ASICs), programmable logic devices (e.g., simple programmable logic devices (SPLDs), complex programmable logic devices (CPLDs), and field programmable gate arrays (FPGAs)), etc.

1. Hardware Configuration In this section, the hardware configuration of this embodiment will be described.

1.1 Information Processing System 1
1 is a configuration diagram showing an information processing system 1 according to this embodiment. The information processing system 1 includes a user terminal 2 and a server 3, which are connected via a network 11. These components will be further described. Here, a system exemplified as the information processing system 1 is one that is made up of one or more devices or components. Therefore, even the server 3 alone is an example of a system.

1.2 User terminal 2
2 is a block diagram showing a hardware configuration of the user terminal 2. The user terminal 2 has a communication unit 21, a storage unit 22, a control unit 23, a display unit 24, and an input unit 25, and these components are electrically connected via a communication bus 20 inside the user terminal 2. Explanations of the communication unit 21, the storage unit 22, and the control unit 23 will be omitted because they are substantially similar to the communication unit 31, the storage unit 32, and the control unit 33 in the server 3 described below.

The display unit 24 may be, for example, included in the housing of the user terminal 2, or may be externally attached. The display unit 24 displays a screen of a graphical user interface (Graphical User Interface: GUI) that can be operated by the user. This is preferably implemented by using display devices such as a CRT display, a liquid crystal display, an organic EL display, and a plasma display depending on the type of the user terminal 2. Here, the display unit 24 will be described as being included in the housing of the user terminal 2.

The input unit 25 may be included in the housing of the user terminal 2, or may be externally attached. For example, the input unit 25 may be implemented as a touch panel integrated with the display unit 24. A touch panel allows the user to input tapping, swiping, and the like. Of course, a switch button, a mouse, a QWERTY keyboard, and the like may be used instead of a touch panel. That is, the input unit 25 accepts operation inputs made by the user. The input is transferred as a command signal via the communication bus 20 to the control unit 23, and the control unit 23 can execute predetermined control or calculations as necessary.

1.3 Server 3
3 is a block diagram showing the hardware configuration of the server 3. The server 3 has a communication unit 31, a storage unit 32, and a control unit 33, and these components are electrically connected via a communication bus 30 inside the server 3. Each component will be further described.

The communication unit 31 is preferably a wired communication means such as USB, IEEE 1394, Thunderbolt, wired LAN network communication, etc., but may also include wireless LAN network communication, mobile communication such as 3G/LTE/5G, Bluetooth (registered trademark) communication, etc. as necessary. In other words, it is more preferable to implement it as a collection of multiple communication means. In other words, the server 3 communicates various information with the user terminal 2 via the communication unit 31 and the network 11.

The memory unit 32 stores various information defined by the above description. This can be implemented, for example, as a storage device such as a solid state drive (SSD) that stores various programs related to the server 3 executed by the control unit 33, or as a memory such as a random access memory (RAM) that stores temporarily required information (arguments, arrays, etc.) related to the program calculations. It may also be a combination of these. In particular, the memory unit 32 stores various programs related to the server 3 executed by the control unit 33.

The control unit 33 processes and controls the overall operations related to the server 3. The control unit 33 is, for example, a central processing unit (CPU) not shown. The control unit 33 realizes various functions related to the server 3 by reading out a specific program stored in the memory unit 32. In other words, information processing by the software stored in the memory unit 32 can be specifically realized by the control unit 33, which is an example of hardware, and executed as each functional unit included in the control unit 33. These will be described in more detail in the next section. Note that the control unit 33 is not limited to being single, and may be implemented with multiple control units 33 for each function. A combination of these may also be used.

2. Functional Configuration In this section, the functional configuration of the present embodiment will be described. As described above, information processing by the software stored in the storage unit 32 is specifically realized by the control unit 33, which is an example of hardware, and can be executed as each functional unit included in the control unit 33.

FIG. 4 is a block diagram showing functions realized by the control unit 33 and the like in the server 3. Specifically, the server 3, which is an example of the information processing system 1, includes an acquisition unit 331, a reference information generation unit 332, a chain processing unit 333, and a display control unit 334.

The acquisition unit 331 is configured to acquire various information necessary for information processing. For example, the acquisition unit 331 may read out various information previously stored in the storage unit 32, or may acquire various information from an external device via the network 11 and the communication unit 31. In particular, the acquisition unit 331 acquires input text KW (input information) received from a user and a preset causal information database IF1 (reference information IF). This will be described in more detail later.

The reference information generation unit 332 is configured to generate reference information IF. More specifically, the reference information generation unit 332 generates the causal information database IF1 (reference information IF) by extracting multiple causal effects based on the text data TX. This will be described in further detail later.

The chain processing unit 333 is configured to generate a causal chain 4 that visually shows the chain of causal relationships. This will be described in more detail later.

The display control unit 334 is configured to generate various display information and control the display content that is visible to the user. The display information may be information itself that is generated in a form that is visible to the user, such as a screen, an image, an icon, text, etc., or may be rendering information for displaying a screen, an image, an icon, text, etc. on the display unit 24 of the user terminal 2. More specifically, the display control unit 334 controls the display of a screen including the related numerical index RI based on the acquired input text KW (input information) and causal information database IF1 (reference information IF). This will be described in more detail later.

3. Information Processing Method In this section, an information processing method of the information processing system 1 described above will be described.

3.1 Generation of Reference Information IF First, the information processing related to the generation of the reference information IF will be described. Fig. 5 is an activity diagram showing the flow of information processing related to the generation of the reference information IF. The reference information IF is information that the server 3 refers to when performing information processing related to the generation of the causal chain 4 (see Fig. 7) described later and the presentation of the related numerical index RI (see Figs. 8 and 9).

For example, the reference information IF may be a causal information database IF1 that includes multiple causal factors that associate causes with effects. Alternatively, the reference information IF may further include information that associates multiple keywords with the associated numerical index RI. According to this aspect, the associated numerical index RI that is associated with the extracted node 4N in the causal chain 4 can be easily presented. Preferably, the associated information is a trained model IF2 that has previously been machine-learned to correlate multiple keywords with the associated numerical index RI.

In addition, the generation of the reference information IF is illustrated as being performed by the server 3. In this case, the user may operate the server 3 directly, or may use the user terminal 2 or the like to operate the server 3 remotely using various methods such as remote desktop. Below, a case will be described in which the user successively generates a causal information database IF1, which is an example of the reference information IF, and a trained model IF2 for presenting the related numerical index RI.

First, a user who wishes to construct a causal information database IF1, which is an example of reference information IF, selects any text data TX. In other words, the acquisition unit 331 in the server 3 acquires at least one piece of text data TX. The text data TX may be, for example, news, topic articles, academic papers, technical specifications, patent documents, research reports, etc. Furthermore, the text data TX is not limited to text files, but may also be a word processing software file, a PDF file, or an HTML file on the Internet (activity A001).

Then, the control unit 33 in the server 3 reads out the dedicated program stored in the memory unit 32 to analyze the text data TX selected in activity A001. As a result, causal pairs relating to various events are extracted from the text data TX (activity A002). There are no particular limitations on the method for extracting causal pairs, and existing natural language processing algorithms may be appropriately adopted.

Then, the control unit 33 in the server 3 reads out the dedicated program stored in the memory unit 32, and adds the causal set extracted in activity A002 to the causal information database IF1, thereby generating the causal information database IF1 (activity A003). In other words, the reference information generation unit 332 generates the causal information database IF1 (reference information IF) by extracting multiple causal effects based on the text data TX. According to this aspect, the causal information database IF1 can be generated according to the user's preferences and the desired usage environment, making it possible to realize prediction of events with more limited uses, etc.

If necessary, the user can select multiple pieces of text data TX. For example, after generating the causal information database IF1 using the text data TX selected the first time, the user can select different text data TX as the second or subsequent processing. In other words, by appropriately repeating activities A001 to A003, a causal information database IF1 with a larger amount of information can be generated.

Next, a user who wants to construct a trained model IF2, which is an example of reference information IF, selects various keywords and numerical indicators that will serve as training data (activity A004). This may be the selection of the text data TX used to extract the causal pairs, or it may be the selection of different data. Next, the user inputs the selected various keywords and numerical indicators into an existing machine learning algorithm, and causes the server 3 to execute machine learning (activity A005). As a result, a trained model IF2 is generated (activity A006). By performing machine learning in this manner, it is possible to present a related numerical indicator RI that is more accurate and relevant than methods that use lookup tables, etc.

If necessary, the user can update the trained model IF2 by repeating machine learning using additional training data. For example, after generating the trained model IF2 using the keywords and numerical indicators selected the first time, further machine learning can be performed using different training data as input for the second or subsequent processing. In other words, by appropriately repeating activities A004 to A006, a trained model IF2 with higher accuracy can be generated.

3.2 Presentation of Causal Chain 4 and Related Numeric Indicator RI Next, a description will be given of information processing related to the presentation of the causal chain 4 and related numerical indicator RI. Fig. 6 is an activity diagram showing the flow of information processing related to the presentation of the causal chain 4 and related numerical indicator RI.

(Overview of the process)
The flow of information processing will be outlined below along each activity in Fig. 6. The user here is assumed to be a person who wants to predict an economic event using the service provided by the information processing system 1. The user, while referring to the screen 5 displayed on the display unit 24 of the user terminal 2, inputs input text KW (input information) related to the event that the user wants to predict using the input unit 25 (activity A101). This input text KW is transmitted to the server 3 via the network 11.

Next, as an acquisition step, the acquisition unit 331 acquires the input text KW (input information) received from the user and the (pre-set) reference information IF generated by the information processing shown in FIG. 5 (activity A102). That is, the control unit 33 searches for a causal pair suitable for the input text KW by matching the input text KW received through the communication unit 31 with the causal information database IF1 stored in the storage unit 32 (activity A103), and determines this (activity A104). Examples of causal pairs will be described later.

When the causal pair is determined, the control unit 33 inputs the keywords included in the causal pair to the trained model IF2 stored in the storage unit 32, and a related numerical index RI that is highly related to the keywords is proposed. Examples of related numerical index RI will be described later.

Next, a screen such as screen 5 is displayed in such a manner that the user can grasp the determined causal set and the proposed related numerical index RI (activity A105). In other words, the control unit 33 reads out a predetermined program stored in the storage unit 32, and extracts the related numerical index RI based on the acquired input text KW (input information) and causal information database IF1 (reference information IF) as an extraction step. Then, the display control unit 334 controls to display screen 5 including the extracted related numerical index RI as a display control step. According to this aspect, the user can input the input text KW related to an event in which the user is interested, and can grasp the related numerical index RI, which is a related numerical index, along with the causal effects of various events related to the input text KW. In other words, the user can obtain a more quantitative prediction result than before.

The user may then input a different input text KW of interest as appropriate. In other words, by repeating activities A101 to A105 as appropriate, the user can obtain prediction results for multiple events.

Furthermore, based on the keywords contained in the causal pair determined in activity A103, it is possible to subsequently search for other causal pairs similar to the one determined in activity A103. Specifically, first, the user uses the input unit 25 to select one of the causal pairs (node 4N) already displayed on the display unit 24 (activity A106).

Then, the control unit 33 reads out the reference information IF (activity A107) and searches for and determines causal pairs related to the selected causal pair by comparing the selected causal pair with the causal information database IF1 stored in the storage unit 32 (activity A108). Examples of related causal pairs will be described later.

In addition, when a related causal pair is determined, the control unit 33 inputs the keywords included in the causal pair into the learned model IF2 stored in the memory unit 32, and a new related numerical index RI that is highly related to the keywords is proposed (activity A109).

Next, the screen such as screen 5 is updated in a manner that allows the user to grasp the determined causal set and the proposed related numerical index RI (activity A110).

That is, by the user appropriately repeating activities A101 to A105 or A106 to A110, a causal chain 4 is generated and displayed so that it can be seen by the user. In other words, the chain processing unit 333, which is a functional unit of the control unit 33, can generate a causal chain 4 that visually shows a chain of causal relationships as a chain processing step. Then, together with the causal chain 4, an associated numerical index RI associated with each causal pair is proposed.

In summary, this information processing method includes the steps of the information processing system 1, and more specifically, the following steps. In the acquisition step, an input text KW (input information) received from a user and a preset causal information database IF1 (reference information) are acquired. The causal information database IF1 (reference information) is information including causality that associates multiple causes with effects. In the extraction step, a related numerical index RI is extracted based on the acquired input text KW (input information) and causal information database IF1 (reference information). The related numerical index RI is related to at least one item that appears in a chain of causal relationships starting from a keyword included in the input text KW (input information).

(Details of information processing)
Next, a detailed description will be given of the information processing outlined in Fig. 6. Fig. 7 is a conceptual diagram for explaining the causal chain 4. Figs. 8 and 9 are diagrams showing an example of the screen 5 of the display unit 24 visually recognized by the user.

As shown in FIG. 7, the causal chain 4 is a network structure having nodes 4N hierarchically, each of which represents a causal set. The causal chain 4 indicates a chain of causal relationships starting from the input text KW entered by the user. More specifically, the causal chain 4 includes multiple nodes 4N that indicate causality in the causal information database IF1. In this example, when the user enters the input text KW, nodes 41, 42, and 43, which are nodes 4N that represent three causal sets, are presented based on the causal information database IF1. As described above, the causal information database IF1 is an example of reference information IF that is preset based on the text data TX.

In addition, in relation to the causal set shown in node 41, nodes 411, 412, and 413 are presented as nodes 4N succeeding node 41. Similarly, in relation to the causal set shown in node 42, nodes 421, 422, and 423 are presented as nodes 4N succeeding node 42, and in relation to the causal set shown in node 43, nodes 431, 432, and 433 are presented as nodes 4N succeeding node 43. According to such a causal chain 4, it is possible to form a causal flow of related events.

Here, it is preferable that the causal effects indicated by the mutually connected nodes 4N are similar to each other. More specifically, the multiple nodes 4N are configured so that the similarity between them can be calculated. The causal chain 4 has a predetermined causal chain 4' including multiple nodes 4N connected in series. For example, in FIG. 7, the input text KW, node 41, and node 411 correspond to the predetermined causal chain 4'. The chain processing unit 333 then adds a new node 4N following the terminal node 4N of the predetermined causal chain 4', and the added new node 4N preferably has a similarity to the terminal node 4N that is equal to or greater than a predetermined value. The concept of similarity is not particularly limited, and the similarity may be determined based on a distance defined in existing natural language processing, for example.

More specifically, the causal chain 4 may be presented on a screen 5 as shown in FIG. 8. In the example of FIG. 8, the user, while referring to the screen 5, inputs "infectious disease" into the input field 50 as the input text KW. When the user presses the search icon 501 in this state, a node 4N indicating a causal pair consisting of a first cause 51 and a first result 52 is displayed in the region R1. The first cause 51 includes the keywords "world situation," "infectious disease," and "impact," and the first result 52 resulting from these includes the keywords "Japanese," "exodus company," "year-on-year comparison," and "ten thousand people."

When the user further selects the keyword "Japanese" 521 in node 4N corresponding to region R1, a node 4N is further displayed in region R2 indicating a causal pair consisting of a second cause 61 and a second result 62. The second cause 61 includes the keywords "Japanese", "guests", and "decrease", and the second result 62 resulting from these includes the keywords "supply and demand", "mitigation", "direction", "guest room", and "unit price".

Although not shown, each time a node 4N is added, a new region and the corresponding node 4N may be displayed under region R2. When adding a subsequent node 4N, the method of selecting the terminal node 4N is not particularly limited. For example, when the user selects keyword "Japanese" 521 in region R1 and presses button 55, a new node 4N is displayed in region R2. In addition, in the example of FIG. 8, starting from the input text KW, a first cause 51 and its result, a first result 52, are presented, and further, a keyword related to the first result 52 is treated as a second cause 61, and a second result 62, its further result, is presented. In other words, a causal chain 4 is presented in the direction of pursuing the result from the cause.

Furthermore, when a node 4N corresponding to the region R2 is selected, a related numerical indicator RI related to the second result 62 is presented together with the causal chain 4. In other words, the related numerical indicator RI is related to at least one item that appears in the chain of causal relationships starting from the input text KW. In this case, the at least one item corresponds to the "guest room". More specifically, the related numerical indicator RI is related to at least one node 4N that appears in the causal chain 4. In addition, when generating such a screen 5, the display control unit 334 controls to display a screen including the causal chain 4 and the related numerical indicator RI. According to this aspect, the related numerical indicator RI is displayed on the screen together with the causal chain 4 in a list format, so that the user can get an overview of the causal flow and the related numerical indicator RI.

Here, when presenting the related association numerical index RI, the learned model IF2 may be referred to. The method of determining the degree of association is not particularly limited, but more specifically, for example, it may be determined based on the modified odds ratio as shown in Equation 1. Moreover, instead of or in addition to the degree of association, the degree of non-association may be determined to present the association numerical index RI.

Preferably, the keywords included in the causal set shown in the node 4N include those indicating large and small increases and decreases, such as "decrease," "increase," "large," and "small," and more preferably, the multiple nodes 4N have polarities that indicate numerical increases or decreases, and the related numerical index RI is determined based on the polarity. In predicting economic events, etc., increasing and decreasing changes are extremely important, so according to this embodiment, the preferred related numerical index RI desired by the user can be presented in consideration of the direction of polarity changes along the causal chain 4 from the input text KW (input information) entered by the user to the related numerical index RI, resulting in a user-friendly specification. Furthermore, it is also possible to implement the method so that not only the related numerical index RI is presented, but also a predicted value of the numerical value can be obtained.

Furthermore, when the user selects the related numerical indicator "room occupancy rate" 63, which is an example of a related numerical indicator RI, while referring to screen 5, it is preferable that the user transitions to screen 5' including a graph of the related numerical indicator RI as shown in FIG. 9. A graph relating to the related numerical indicator "room occupancy rate" 63 is displayed on screen 5'. By allowing the user to refer to such a graph, the user can more intuitively understand the prediction results.

In this manner, in this embodiment, an information processing system 1 is proposed that performs a chain search for causal relationships of economic events, etc. from a causal information database IF1 extracted from text data TX, and presents a related numerical index RI related to ripple effects. According to the information processing method using such information processing system 1, users can more easily obtain quantitative prediction results regarding various economic events, etc.

4. Others The information processing system 1 according to the present embodiment may be modified in the following manner.

In the above embodiment, the configuration of the information processing system 1 has been described, but a program that causes a computer to execute each step of the information processing system 1 may be provided. Such a program may be configured to be executable offline on a stand-alone computer that is separated from the network 11.

When presenting the related numerical index RI, rather than just presenting the index itself, it may be implemented to provide a forecast of time-dependent changes, such as an upward change in the value in X days or within X months.

At least one text data TX selected in generating reference information such as the causal information database IF1 may be a plurality of text data TX written in different languages. For example, the first extraction of causal pairs may be performed using text data TX written in Japanese, and the second extraction of causal pairs may be performed using text data TX written in a foreign language such as English or Chinese. In such a case, the chain processing unit 333 generates a causal chain 4 that visually shows a chain of causal relationships, and the causal chain 4 includes a plurality of nodes 4N that show causality described in different languages. In other words, even in different languages, by extracting common concepts or keywords with the same meaning, it is possible to compare the similarity of the nodes 4N extracted in each different language. According to such an embodiment, a more universal causal chain 4 that transcends language barriers can be generated.

In this embodiment, a new node 4N related to the terminal node 4N is searched for and determined based on the causal information database IF1. However, other nodes 4N may be referenced in addition to the terminal node 4N. For example, other nodes 4N located in parallel in the same hierarchy, such as node 411 or node 413 as seen from node 412, may be referenced. However, it is particularly preferable that the chain processing unit 333 weights multiple nodes 4N in a given causal chain 4' and determines a new subsequent node 4N based on the weighted evaluation value. According to this aspect, a new node 4N that follows the previous causal flow is proposed in addition to the immediately preceding node 4N, allowing the user to make more accurate predictions.

In this embodiment, particularly in FIG. 8, the causal chain 4 is presented in a direction from the cause to the result, but it is also possible to have the user input input text KW related to the result, and present the causal chain 4 in a direction from the result to the cause. FIGS. 10 and 11 are diagrams showing examples of different aspects of the screen 7 of the display unit 24 that is visually recognized by the user.

In FIG. 10, the user, while referring to screen 7, inputs "infectious disease" as the input text KW into input field 70. When the user presses search icon 701 in this state, nodes 81, 82, and 83, as well as multiple nodes 4N, are presented to the right of input field 70. When the user presses button 751, nodes 821, 822, and 823 are added to the right, as shown in FIG. 11. The added nodes 821, 822, and 823 are connected by curved lines so that it is visible that they are nodes 4N following node 82. When the user presses button 752, other nodes 4N that are in a parallel relationship with nodes 81, 82, and 83 may be displayed.

Here, as shown in FIG. 10 and FIG. 11, the screen 7 has a selection button 76 and a selection button 77 below the input field 70. When the user selects the selection button 76, the causal chain 4 can be displayed in a direction from the cause to the result. Although not shown, when the user selects the selection button 77, the causal chain 4 can be displayed in a direction from the result to the cause. In general, since multiple results occur due to one item as a cause, the presentation of the direction from the cause to the result is considered to be a forward problem. Conversely, the presentation of the direction from the result to the cause is an inverse problem, and in order to perform more accurate cause prediction, it is preferable to set a separate constraint condition in addition to the judgment of similarity. Such constraint conditions are not particularly limited, and it is possible to appropriately assume that the weighting value is changed or the subsequent node 4N is restricted.

In this embodiment, the causal chain 4 can theoretically be added almost without limit as long as the information in the causal information database IF1 is enriched, but it is expected that as the number of nodes 4N increases, the relevance to the input text KW initially entered by the user will become weaker. To prevent this, the number of layers of nodes 4N that can be included in one causal chain 4 may be limited. The number of layers to be limited is not particularly limited, but specifically, for example, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 layers, or may be within a range between any two of the values exemplified here. More preferably, 1 to 5 layers, and particularly preferably, 3 layers.

In this embodiment, an example has been described in which a causal chain 4 and a related numerical index RI are generated starting from the input text KW and are presented, but the related numerical index RI may be used as an input and a causal chain 4 including a related causal node 4N may be presented. In other words, the acquisition unit 331 acquires a numerical index (related numerical index RI) received from the user and preset reference information IF as an acquisition step. The reference information IF is information that associates multiple keywords with numerical indexes, such as a trained model IF2. Then, the control unit 33 extracts a causal chain 4, which is a chain of causal relationships including keywords, based on the acquired numerical index and reference information as an extraction step.

Furthermore, it may be provided in the following aspects:
The information processing system is further configured to execute a display control step, in which a screen including the extracted related numerical indicator is controlled to be displayed.
In the information processing system, the reference information further includes information associating a plurality of keywords with the associated numerical indicator.
In the information processing system, the associated information is a trained model that has been machine-learned in advance to determine the correlation between the multiple keywords and the related numerical indicators.
The information processing system further includes a chain processing step, in which a causal chain is generated that visually shows a chain of causal relationships, the causal chain including a plurality of nodes that indicate causality in the reference information, and the display control step controls to display a screen including the causal chain and the associated numerical indicator, the associated numerical indicator being related to at least one node that appears in the causal chain.
In the information processing system, the multiple nodes are configured to be able to calculate similarities between each other, the causal chain includes a predetermined causal chain including multiple nodes arranged in series, and in the chain processing step, a new node is added following the terminal node of the predetermined causal chain, and the new node added here has a similarity with the terminal node that is equal to or greater than a predetermined value.
In the information processing system, the chain processing step weights a plurality of nodes in the predetermined causal chain, and determines a new subsequent node based on the weighted evaluation value.
In the information processing system, the plurality of nodes have a polarity representing a numerical increase or decrease, and the associated numerical indicator is determined based on the polarity.
The information processing system further includes a reference information generating step, in which the acquisition step acquires at least one text data, and the reference information generating step generates the reference information by extracting multiple causalities based on the text data.
The information processing system further includes a chain processing step, wherein the at least one text data is a plurality of text data written in different languages, and the chain processing step generates a causal chain that visually shows a linked causal relationship, wherein the causal chain includes a plurality of nodes that indicate causality each described in the different languages.
An information processing system configured to execute the following steps: in an acquisition step, a numerical indicator received from a user and preset reference information are acquired, the reference information being information that associates a plurality of keywords with the numerical indicator; and in an extraction step, a chain of causal relationships including the keywords are extracted based on the acquired numerical indicator and the reference information.
An information processing method comprising the steps of the information processing system.
A program that causes a computer to execute each step of the information processing system.
Of course, this is not the case.

Finally, although various embodiments of the present invention have been described, these are presented as examples and are not intended to limit the scope of the invention. The novel embodiments can be embodied in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. The embodiments and their modifications are within the scope and spirit of the invention, and are included in the scope of the invention and its equivalents as set forth in the claims.

1: Information processing system 11: Network 2: User terminal 20: Communication bus 21: Communication unit 22: Memory unit 23: Control unit 24: Display unit 25: Input unit 3: Server 30: Communication bus 31: Communication unit 32: Memory unit 33: Control unit 331: Acquisition unit 332: Reference information generation unit 333: Chain processing unit 334: Display control unit 4: Causal chain 4': Predetermined causal chain 4N: Node 41: Node 411: Node 412: Node 413: Node 42: Node 421: Node 422: Node 423: Node 43: Node 431: Node 432: Node 433: Node 5: Screen 5': Screen 50: Input field 501: Search icon 51: First cause 52: First result 521: Keyword "Japanese"
55: Button 61: Second cause 62: Second result 63: Related numerical indicator "room occupancy rate"
7: Screen 70: Input field 701: Search icon 751: Button 752: Button 76: Selection button 77: Selection button 81: Node 82: Node 821: Node 822: Node 823: Node 83: Node IF: Reference information IF1: Causal information database IF2: Trained model KW: Input text R1: Region R2: Region RI: Related numerical index TX: Text data

Claims (12)

An information processing system,
A control unit is provided,
The control unit is configured to execute the following steps:
In the acquisition step, input information received from a user and preset reference information are acquired;
Here, the reference information is information indicating a causal relationship that associates a plurality of items indicating causes with items indicating results , and further includes information that associates a plurality of keywords with related numerical indicators,
Here, the related numerical index is a numerical index related to at least one of the items indicating the cause and the items indicating the result appearing in the chain of causal relationships starting from a keyword included in the input information,
In the extraction step, the system extracts the relevant numerical indicators based on the acquired input information and the acquired reference information. 2. The information processing system according to claim 1 ,
The system, wherein the associated information is a trained model that has been machine-learned to determine the correlation between the multiple keywords and the related numerical indicators. 3. The information processing system according to claim 1,
Furthermore, in the display control step, the system displays a screen including the extracted related numerical indicators. 4. The information processing system according to claim 3 ,
Furthermore, in the chain processing step, a causal chain is generated that visually indicates the linked causal relationships, and the causal chain includes a plurality of nodes that respectively indicate the causal relationships in the reference information;
In the display control step, a screen including the causal chain and the associated numerical indicators is displayed,
A system in which the relevant numerical indicator is related to the causal relationship indicated by at least one node appearing in the causal chain. 5. The information processing system according to claim 4 ,
The plurality of nodes are configured to be able to calculate a similarity between each other,
the causal chain comprises a predetermined causal chain including a plurality of nodes in series;
In the chain processing step, a new node is added that follows the terminal node of the specified causal chain, and the new node added here has a similarity to the terminal node that is equal to or greater than a predetermined value. 6. The information processing system according to claim 5 ,
In the chain processing step, the system weights the similarity between each of the new node and multiple nodes serially connected in the specified causal chain , and determines the subsequent new node based on the weighted similarity . In the information processing system according to any one of claims 4 to 6 ,
the plurality of nodes have polarities representing numerical increase or decrease;
A system in which the related numerical indicator is determined to be an indicator suitable for predicting events of the directional numerical change indicated by the polarity of the node indicating the causal relationship to which the related numerical indicator is related . In the information processing system according to any one of claims 4 to 7,
In the display control step, a list including a plurality of the related numerical indicators is displayed, and when any one of the related numerical indicators is selected from the list, a graph relating to the related numerical indicator is displayed. In the information processing system according to any one of claims 1 to 8,
In the acquiring step, at least one piece of text data is acquired,
Furthermore, in the reference information generating step, the system generates the reference information by extracting a plurality of causalities based on the text data. 10. The information processing system according to claim 9,
The at least one text data is a plurality of text data written in different languages,
Further, in the chain processing step, the system generates a causal chain that visually represents a chain of causal relationships, wherein the causal chain includes a plurality of nodes that represent causal effects each described in the different languages. 1. An information processing method, comprising:
A method comprising the steps of the information processing system according to any one of claims 1 to 10 , wherein each of the steps is executed by a control unit of the information processing system . A program,
A program for causing a computer to execute each step of the information processing system according to any one of claims 1 to 10 .