JP7488207B2 - Future event estimation system and future event estimation method - Google Patents

Description

The present invention relates to a future event estimation system and a future event estimation method.

Non-Patent Document 1 describes a technology that automatically acquires knowledge of relationships between events, such as causal relationships, from large-scale text data. Here, an event refers to an action that an actor intentionally performs and other occurrences (experiences, states, changes in state, etc.). Furthermore, a relationship between events refers to, for example, a relationship in which the event "putting in the laundry" is likely to occur after the event "it's raining" occurs.

In light of the current trends, such as advances in statistical language processing and the availability of large volumes of documents describing human behavior, the above technology seeks to acquire knowledge of inter-event relationships from text corpora using co-occurrence patterns of two event expressions, such as "doing ~ in order to ~." The above technology also uses event-implying nouns that directly or indirectly refer to events, and considers applying co-occurrence pattern acquisition technology for extracting entity relationships to the acquisition of inter-event relationships, thereby preparing a large number of special co-occurrence patterns with strong semantic restrictions and aiming to achieve both precision and scale in knowledge acquisition.

Shuya Abe, Kentaro Inui, Yuji Matsumoto, "Acquisition of knowledge of inter-event relations using event-implicative nouns", Information Processing Society of Japan, 2006/11/23, pp.95-100

In order to make appropriate decisions regarding the measures that the organization should take, organizations such as companies and government agencies are required to accurately predict events that may actually occur in the future (hereafter referred to as "future events") based on current events (hereafter referred to as "current events"). For example, if it were possible to predict a future event such as "Japan will impose export restrictions on automobiles" in advance when a current event such as "Japan-U.S. trade talks will be held" is occurring, the government could take measures in advance such as "considering measures to stimulate the economy in preparation for a recession throughout Japan," and companies could take measures in advance such as "holding back on procurement of parts in preparation for reduced automobile production."

By the way, future events are not only determined by current events, but also change depending on events related to the current event (hereafter referred to as "peripheral events"). For example, if the current event of "Japan-US trade talks in 1981" has the peripheral event of "the US trade deficit is excessive," then the future event of "Japan will impose restrictions on automobile exports" will be predicted. On the other hand, if the current event of "Japan-US trade talks in 2018" has the peripheral event of "US tariffs are also being viewed as problematic," then the future event of "US will lower tariffs on automobiles" will be predicted.

In the above-mentioned non-patent document 1, verb phrases and event nouns are acquired as references to events, but descriptions of events surrounding the event are ignored. Therefore, while it is possible to identify causal relationships that exist without relying on surrounding events, such as "when it rains, the laundry is brought in," it is not necessarily possible to accurately predict future events when surrounding events need to be taken into account.

The present invention has been made against this background, and aims to provide a future event estimation system and a future event estimation method that are capable of predicting future events with high accuracy.

One of the present inventions for achieving the above-mentioned object is a future event inference system, which is configured using an information processing device having a processor and a memory, and includes a memory unit that stores event-related mention information, which is information regarding event-related mentions, which are mentions of events included in document information, a user mention extraction unit that extracts user mentions from the event-related mention information, which are mentions related to event designation information, which is information that designates a specific event, and a future mention identification unit that identifies future mentions from the event- related mention information, which are mentions of future events regarding the event represented by the user mention, wherein the user mention extraction unit extracts a plurality of the user mentions from the event-related mention information and generates a user mention set that is a set whose elements are the extracted plurality of user mentions, and the future mention identification unit extracts a plurality of mentions that are candidates for the future mentions from the event-related mention information as candidates, generates a plurality of future mention sets that are sets obtained by classifying the extracted candidates, and calculates the causal strength between the user mention set and each of the plurality of future mention sets .

Problems, configurations, and advantages other than those described above will become clear from the description of the embodiment of the invention below.

The present invention makes it possible to predict future events with high accuracy.

1 is a diagram showing a schematic configuration of a future event estimation system according to a first embodiment; 1 is an example of an information processing device used in the configuration of a future event estimation system. 1 is an example of document information. 13 is an example of event-related mention information. 13 is an example of an event specification information setting screen. This is an example of future mention sets and causal strength. 13 is a flowchart illustrating a future mention information generating process. 13 is an example of a future event information presentation screen. FIG. 11 is a diagram showing a schematic configuration of a future event estimation system according to a second embodiment. 13 is an example of an event designation information setting screen according to the second embodiment. 13 is an example of a future event information presentation screen displayed by the future event estimation system of the third embodiment.

The following describes in detail an embodiment of the present invention with reference to the drawings. Note that the following description and drawings are merely examples for explaining the present invention, and have been omitted or simplified as appropriate for clarity of explanation. The present invention can also be implemented in various other forms. Furthermore, unless otherwise specified, each component may be singular or plural.

In the following explanation, the same or similar components may be given the same reference numerals and duplicate explanations may be omitted. In addition, in the following explanation, the letter "S" before a reference numeral denotes a processing step. Furthermore, various information (data) in the following explanation may be expressed or managed in a manner other than the data structure shown as an example. In addition, in the following explanation, when explaining various types of identification information, expressions such as "identifier" and "ID" are used as appropriate, but these are interchangeable.

[First embodiment]
1 shows a schematic configuration of an information processing system (hereinafter, referred to as a "future event inference system 1") according to a first embodiment. The future event inference system 1 is configured using one or more information processing devices (computers).

The future event estimation system 1 acquires document information from a user or from an external source such as another information processing system, extracts mentions of events (hereinafter referred to as "event-related mentions") from the acquired document information, and manages information related to the extracted event-related mentions as event-related mention information. The event-related mention information includes information related to events related to the event-related mentions (hereinafter referred to as "peripheral events").

The future event estimation system 1 accepts input of information (hereinafter referred to as "event specification information") that specifies a specific event (e.g., an event that is currently occurring (hereinafter referred to as "current event")) from a user, and extracts mentions (hereinafter referred to as "user mentions") related to the accepted event specification information from the event-related mention information. The user is, for example, a person in charge of analyzing events in an organization such as a company or a government agency.

The future event estimation system 1 identifies from the event-related mention information a mention of a future event (hereinafter referred to as a "future mention") regarding an event represented by a user mention of a current event, etc., generates information (hereinafter referred to as "future event information") regarding an event that may occur in the future (hereinafter referred to as a "future event") based on the identified future mention, and provides the generated future event information to the user.

FIG. 2 shows an example of a hardware configuration of an information processing device used in the configuration of the future event estimation system 1. The illustrated information processing device 10 includes a processor 11, a main memory device 12, an auxiliary memory device 13, an input device 14, an output device 15, and a communication device 16. The illustrated information processing device 10 may be realized, in whole or in part, by using virtual information processing resources provided by a virtualization technology or a process space separation technology, such as a virtual server provided by a cloud system. In addition, all or in part of the functions provided by the information processing device 10 may be realized by a service provided by a cloud system via an API (Application Programming Interface), for example. The future event estimation system 1 may be configured by using a plurality of information processing devices 10 connected to each other so as to be able to communicate with each other.

In the figure, the processor 11 is, for example, a CPU (Central Processing Unit),
PU (Micro Processing Unit), GPU (Graphics Processing Unit), FPGA (Field Programmable Gate Array), ASIC (Application Specific Integrated Circuit
), AI (Artificial Intelligence) chips, etc.

The main memory device 12 is a device for storing programs and data, and is, for example, a ROM (Read Only Memory).
Only Memory), RAM (Random Access Memory), NVRAM (Non Volatile RAM), etc.

The auxiliary storage device 13 is, for example, a solid state drive (SSD), a hard disk drive, an optical storage device (CD (Compact Disc), DVD (Digital Versatile Disc), etc.), a storage system, an IC card, a read/write device for a recording medium such as an SD card or an optical recording medium, a storage area of a cloud server, etc. Programs and data can be read into the auxiliary storage device 13 via a recording medium reader or a communication device 16. The programs and data stored (memorized) in the auxiliary storage device 13 are read into the main storage device 12 as needed.

The input device 14 is an interface that accepts input from the outside, and is, for example, a keyboard, a mouse, a touch panel, a card reader, a pen-input tablet, a voice input device, etc.

The output device 15 is an interface that outputs various information such as the progress of processing and the results of processing. The output device 15 is, for example, a display device (liquid crystal monitor, LCD (Liquid Crystal Display), graphic card, etc.) that visualizes the various information described above, a device that converts the various information described above into audio (audio output device (speaker, etc.)), or a device that converts the various information described above into text (printer, etc.). Note that, for example, the information processing device 10 may be configured to input and output information to and from other devices via the communication device 16.

The input device 14 and output device 15 constitute a user interface that accepts information from the user and presents information.

The communication device 16 is a device that realizes communication with other devices. The communication device 16 is a wired or wireless communication interface that realizes communication with other devices via the communication network 5, and is, for example, a NIC (Network Interface Card), a wireless communication module, a USB module, etc.

The information processing device 10 may be equipped with, for example, an operating system, a file system, a DBMS (DataBase Management System) (relational database, NoSQL, etc.), a KVS (Key-Value Store), etc.

Each function of the future event inference system 1 is realized by the processor 11 reading and executing a program stored in the main memory device 12, or by the hardware (FPGA, ASIC, AI chip, etc.) that constitutes the future event inference system 1. The future event inference system 1 stores the various information (data) described above, for example, as a database table or a file managed by a file system.

As shown in FIG. 1, the future event inference system 1 includes the functions of a memory unit 110, a document information acquisition unit 120, an event-related mention information generation unit 130, an event specification information reception unit 140, a future event inference unit 150, and a future event information generation unit 160.

Of the above functions, the memory unit 110 stores document information 111, event-related mention information 112, event specification information 113, user mentions 114, future mention information 115, and future event information 116.

The document information acquisition unit 120 acquires the document information 111. In this example, the document information 111 is assumed to be data in a text data format. Furthermore, in this embodiment, the information (data) acquired from the document information 111 is also managed as text data in principle. However, the data format of the document information 111 and the information (data) acquired from the document information 111 is not necessarily limited. It is sufficient that this information (data) is in a data format that can at least be handled by the information processing device 10.

The document information acquisition unit 120 acquires the document information 111 by communicating with another information processing device via, for example, a communication network (the Internet, a local area network (LAN), a wide area network (WAN), a dedicated line, a public communication network, etc.). The document information acquisition unit 120 also acquires the document information 111 from a user via, for example, a Web Application Programming Interface (Web API). The document information acquisition unit 120 also acquires the document information 111 from a user via, for example, a Web Application Programming Interface (Web API).
The document information acquisition unit 120 acquires the document information 111, for example, by distributing news articles using RSS (Really Simple Syndication, Rich Site Summary) or automatically collecting the document information by web crawling. The document information acquisition unit 120 may provide the user with a function for editing the document information 111 (such as a function for adding supplementary information such as the writing time of the document) via a user interface.

Figure 3 shows an example of document information 111. The example document information 111 includes one or more entries (records) having the following items: document ID 1111, content 1112, and time information 1113.

A document ID that is an identifier for a document is set in document ID 1111. The type of document ID is not necessarily limited, and may be, for example, the title of an article or book in which the document is written, or information indicating the location of the document (for example, a URL (Uniform Resource Locator)).

Content 1112 is set to the content (text data) of the document identified by the document ID.

In the time information 1113, information (hereinafter referred to as "time information") that can identify the time (date, time period, etc.) when the document or the information described in the document occurred is set. The time information 1113 is attached to the document information 111, for example, as meta information. Note that the time information does not necessarily need to indicate the exact time when the document was created, but can at least be information that can identify the time when the event described in the document occurred (date and time of writing, date and time of publication in a book such as a magazine, the publication date of the book, date and time when the document was made public on the Internet, etc.).

Returning to FIG. 1, the event-related mention information generation unit 130 extracts event-related mentions from the document information 111 to generate event-related mention information 112.

Figure 4 shows an example of event-related mention information 112. The illustrated event-related mention information 112 includes one or more entries (records) each consisting of the following items: mention ID 1121, event-related mention information 1122, surrounding event mention 1123, time information 1124, and document ID 1125. One entry of the event-related mention information 112 corresponds to one event-related mention.

Of the above items, mention ID 1121 is set to an identifier for an event-related mention (hereinafter referred to as the "mention ID").

Event-related mention information 1122 is set with an event-related mention. For example, the event-related mention information generating unit 130 extracts, from the document information 111, a character string combining a predicate identified by predicate-argument structure analysis and an identified phrase (argument) as an event-related mention. The event-related mention information generating unit 130 also extracts, from the document information 111, an event-related mention identified by syntactic analysis such as identifying a noun phrase or a verb phrase. The event-related mention information generating unit 130 also extracts, from the document information 111, a word or phrase registered in advance by the user or a word or phrase related to the word or phrase (such as a similar word or phrase) as an event-related mention.

In the peripheral event mention 1123, a mention related to a peripheral event of the event-related mention (hereinafter, referred to as a "peripheral event mention"). The event-related mention information generating unit 130 extracts peripheral events by, for example, selecting other words and phrases included in the document including the event-related mention from the document information 111. The event-related mention information generating unit 130 performs the above selection by, for example, a method of prioritizing selection of words and phrases with high frequency, or a method of selecting the top n words and phrases with high TF-IDF (Term Frequency - Inverse Document Frequency) values (words with high importance). The event-related mention information generating unit 130 also performs the above selection by, for example, selecting all words and phrases in the document information 111.

The content of the time information 1113 of the document information 111 corresponding to the event-related mention is set in the time information 1124. When information related to time is included in the document information 111 including the mention or in meta-information associated with the document information 111, such information is set in the time information 1124.

The document ID 1125 is set to the document ID 1111 of the document information 111 from which the event-related mention was obtained.

Returning to FIG. 1, the event specification information receiving unit 140 receives event specification information from the user via the user interface.

Figure 5 is an example of a screen (hereinafter referred to as "event designation information setting screen 500") that the event designation information receiving unit 140 presents to the user when receiving event designation information. The event designation information receiving unit 140 receives input of event designation information 113 from the user via the event designation information setting screen 500. The event designation information setting screen 500 may be presented to the user, for example, by a GUI (Graphical User Interface) such as a web page.

As shown in the figure, the event specification information setting screen 500 has input fields for event 511, peripheral event 512, and time information 513. The event specification information setting screen 500 also has a search button 520. Note that the event specification information setting screen 500 may include input fields for other supplementary information in addition to the input fields shown as examples. When the user sets content in each input field and operates the search button 520, the event specification information receiving unit 140 generates information (dataset) that associates the content set in each input field of the event specification information setting screen 500 as event specification information 113.

Returning to FIG. 1, as shown in the figure, the future event estimation unit 150 has a user mention extraction unit 151 and a future mention identification unit 152.

The user mention extraction unit 151 extracts mentions (user mentions 114) related to the event designation information 113 received by the event designation information receiving unit 140 via the event designation information setting screen 500 from the event-related mention information 112. The future mention identification unit 152 also identifies mentions of future events about the event represented by the user mention 114 from the event-related mention information 112 to generate future mention information 115. The functions of the user mention extraction unit 151 and the future mention identification unit 152 are described in detail below.

First, a description will be given of the user mention extraction unit 151. The user mention extraction unit 151 extracts mentions related to the event designation information 113 from the event-related mention information 112, and generates a user mention set Q (user mentions 114) expressed by the following formula.

For example, the user mention extraction unit 151 calculates the similarity S _i between the event designation information 113 and each mention m _i in the event-related mention information 112, and extracts a predetermined number of mentions (or mentions exceeding a preset threshold) from the event-related mention information 112 based on the calculated similarity S _i , and sets these as elements of the user mention set Q.

The method of calculating the similarity S _i is not necessarily limited. For example,
The similarity S _i is calculated by the following method, which will be exemplified below, depending on which item of the event specification information 113 is used, what scale is used as the scale of similarity, etc.

For example, the user mention extraction unit 151 calculates the similarity S _i between the string in the event-related mention and the string in the mention m _i by applying the string similarity S, which is a similarity based on the similarity of strings. The string similarity S is calculated from the following formula using, for example, the distance between strings a and b, d (string a, string b), and a function L that takes a string as input and returns the string length.

The user mention extraction unit 151 uses the obtained string similarity S to obtain a similarity S _i , for example, from the following formula.

Furthermore, for example, the user mention extraction unit 151 calculates the similarity S _i based on the character string representing the peripheral event and the character string of the peripheral event mention in the user mention m _i using the following formula.

In addition, for example, the user mention extraction unit 151 extracts the time information and the mention m _i
The similarity S _i is calculated by applying the time similarity S _{t ,} which indicates the closeness in ^time to the time information specified by ...

The user mention extraction unit 151 uses the time similarity S _t to calculate the similarity S _i from, for example, the following formula.

The user mention extraction unit 151 may calculate the similarity by combining two or more of the above-described methods for calculating the similarity _S. In this case, for example, the similarity _S calculated by each method may be weighted and summed to obtain the final similarity.

Next, the future mention identification unit 152 shown in Fig. 1 will be described. The future mention identification unit 152 uses the user mention set Q (user mentions 114) generated by the user mention extraction unit 151 to extract a group of mentions that are likely to indicate the same future event from the event-related mention information 112, and generates a future mention set F shown in the following formula.

The future mention identification unit 152 also determines the strength of the causal relationship (hereinafter referred to as the "causal strength score S(F,Q)") between the future event indicated by the future mention set F and the event indicated by the user mention set Q. Then, the future mention identification unit 152 generates future mention information 115 based on the future mention set F and the causal strength score S(F,Q).

Figure 6 is an example of future mention information 115. As shown in the figure, future mention information 115 is composed of one or more entries (records) having each of the fields of future mention set ID 1151, future mention ID 1152, and causal strength 1153. One entry of future mention information 115 corresponds to one future mention set F.

Of the above items, the future mention set ID 1151 is set with an identifier of the future mention set F (hereinafter, referred to as the "future mention set ID"). The future mention ID 1152 is set with a mention ID of a mention that is an element of the future mention set F. The causal strength 1153 is set with a causal strength score S(F,
Q) is set.

Figure 7 is a flowchart explaining the process performed by the future mention identification unit 152 (hereinafter referred to as "future mention information generation process S700"). The future mention information generation process S700 will be explained below with reference to this figure.

First, the future mention identification unit 152 extracts mentions that are candidates for future mentions from the event-related mention information 112 as candidates (S711). For example, the future mention identification unit 152 extracts all mentions in the event-related mention information 112 as candidates. Also, for example, the future mention identification unit 152 extracts mentions that have the same document ID (associated as the source) as the document from which the elements (user mentions) of the user mention set Q were obtained, from the event-related mention information 112, as candidates.

Next, the future mention identification unit 152 identifies mentions that are estimated to represent the same or similar events for the extracted candidates, and generates a future mention set with the identified mentions as elements (S712). For example, the future mention identification unit 152 classifies the extracted candidates into multiple clusters that are estimated to represent the same or similar events by using a similarity matrix related to the string similarity S between the mention strings and a clustering method based on a similarity matrix such as spectral clustering, and generates multiple future mention sets F with each cluster obtained by classification as a future mention set F.

Next, the future mention identification unit 152 calculates a causal strength score S(F,Q) between the events indicated by each of the generated multiple future mention sets F and the events represented by the event designation information 113.
For example, the future mention identification unit 152 calculates the inter-document pointwise mutual information (PMI) between the future mention set F and the user mention set Q (S713).
is calculated and used as the causal strength score S(F,Q).

Next, the future mention identification unit 152 generates future mention information 115 based on the information generated as described above (S714).

Returning to FIG. 1, the future event information generation unit 160 shown in the figure generates future event information 116 based on the future mention information 115 generated by the future event estimation unit 150 as described above, and presents the contents of the generated future event information 116 to the user.

8 shows an example of a screen (hereinafter referred to as a "future event information presentation screen 800") that the future event information generating unit 160 displays via the output device 15 when presenting the contents of the future event information 116 to the user. The future event information generating unit 160 presents the future event information presentation screen 800 to the user as a Web page, for example. The future event information presentation screen 800 may be a common screen with the event designation information setting screen 500.

In the illustrated future event information presentation screen 800, the elements (future mentions) of the future mention set F are displayed for each future mention set F in descending order of causal strength 1153. By referring to the future event information presentation screen 800, the user can grasp the future events corresponding to the input event specification information, along with a priority order indicating the likelihood of occurrence.

In addition, since the user mention set Q has elements that are mentions extracted from the event-related mention information 112 as being related to the event designation information 113, the causal strength score S(F,Q) also reflects the similarity in time (time information of the elements of each set) between the user mention set Q and the future mention set F. In other words, if there is a time difference between the two, the causal strength score S(F,Q) becomes smaller, and it is possible to prevent information with a large time difference from the user mention from being extracted as a future mention, thereby improving the estimation accuracy of the future event.

[Second embodiment]
9 shows a schematic configuration of a future event inference system 1 shown as the second embodiment. The configuration of the future event inference system 1 of the second embodiment is basically the same as that of the future event inference system 1 of the first embodiment, but the future event inference system 1 of the second embodiment further includes an event designating information presenting unit 135 and an event designating information candidate extracting unit 170, and the storage unit 110 further stores event designating information candidate information 117.

When the event designation information receiving unit 140 receives the event designation information 113 from the user, the event designation information presentation unit 135 presents on the output device 15 information acquired from the event designation information candidate information 117 in a selectable state as a candidate for the event designation information 113. This eliminates the need for the user to set the contents of the event designation information 113 one by one via the event designation information setting screen 500 as in the first embodiment, and allows the user to efficiently input the event designation information 113. Note that the above candidates may be, for example, information extracted from the event designation information candidate information 117 using a keyword received from the user. The above candidates may also be all information acquired from the event designation information candidate information 117.

The event designation information candidate extraction unit 170 generates the event designation information candidate information 117 based on, for example, the event-related mention information 112. By generating the event designation information candidate information 117 based on the event-related mention information 112, the future event estimation unit 150 can accurately generate mentions that are candidates for the event designation information 113. The event designation information candidate information 117 may be common to the event-related mention information 112, for example. In addition, the event designation information candidate extraction unit 170 may generate the event designation information candidate information 117 based on, for example, information received from a user.

Figure 10 shows an example of an event designation information setting screen 500 presented to the user by the event designation information receiving unit 140 of the future event inference system 1 of the second embodiment. In the illustrated event designation information setting screen 500, a plurality of pieces of information (multiple sets consisting of the items of event 511, surrounding event 512, and time information 513 and a search button 520) acquired from the event designation information candidate information 117 are displayed as candidates for the event designation information 113. When the user operates the search button 520 for the information to be selected, the future event inference unit 150 generates future mention information 115 based on the corresponding event designation information 113. Next, the future event information generation unit 160 generates future event information 116, and generates a future event information presentation screen 800 based on the generated future event information 116 to present to the user.

[Third embodiment]
The future event inference system 1 of the third embodiment basically has the same configuration as the future event inference system 1 of the first embodiment. However, the future event inference system 1 of the third embodiment differs in configuration from the future event inference system 1 of the first embodiment in that when the future event information generation unit 160 displays the future mention set F on the future event information presentation screen 800, it highlights representative future mentions in each future mention set F.

For example, the future event information generating unit 160 identifies as a representative future mention a mention having the earliest time (date, time, etc.) identified from each time information among the respective mentions for each future mention set F. In addition, for example, the future event information generating unit 160 identifies as a representative future mention a mention having the largest sum of similarities with other mentions belonging to the same future mention set F among the respective mentions for each future mention set F. In this case, the similarity is obtained, for example, by applying string similarity S to the mention strings or surrounding event strings in the mention, or by applying time similarity S _t to the time information.

Figure 11 shows an example of a future event information presentation screen 800 displayed by the future event estimation system 1 (future event information generation unit 160) of the third embodiment. As shown in the figure, the future event information generation unit 160 displays the most representative future mentions among the future mentions for each future mention set F in bold, thereby distinguishing them from the other future mentions. By highlighting the most representative future mentions in this way, the user can accurately grasp the future events.

Although one embodiment of the present invention has been described above, it goes without saying that the present invention is not limited to the above embodiment and can be modified in various ways without departing from the spirit of the invention. For example, the above embodiment has been described in detail to clearly explain the present invention, and is not necessarily limited to having all of the configurations described. In addition, it is possible to add, delete, or replace part of the configuration of the above embodiment with other configurations.

In addition, the above-mentioned configurations, functional units, processing units, processing means, etc. may be realized in part or in whole by hardware, for example, by designing them as integrated circuits. In addition, the above-mentioned configurations, functions, etc. may be realized in software by a processor interpreting and executing a program that realizes each function. Information such as the program, table, file, etc. that realizes each function can be stored in a storage device such as a memory, a hard disk, or an SSD (Solid State Drive),
It can be placed on a recording medium such as an IC card, an SD card, or a DVD.

Furthermore, the layout of the various functional units, various processing units, and various databases of each information processing device described above is merely an example. The layout of the various functional units, various processing units, and various databases can be changed to an optimal layout in terms of the performance, processing efficiency, communication efficiency, etc. of the hardware and software equipped in these devices.

In addition, the configuration (schema, etc.) of the database that stores the various types of data mentioned above can be flexibly changed from the perspective of efficient use of resources, improved processing efficiency, improved access efficiency, improved search efficiency, etc.

1 Future event estimation system, 110 Storage unit, 111 Document information, 112 Event-related mention information, 113 Event designation information, 114 User mention, 115 Future mention information, 116 Future event information, 117 Event designation information candidate information, 120 Document information acquisition unit, 130 Event-related mention information generation unit, 135 Event designation information presentation unit, 140 Event designation information reception unit, 150
Future event estimation unit, 151 user mention extraction unit, 152 future mention identification unit, 160 future event information generation unit, 170 event designation information candidate extraction unit, 500 event designation information setting screen, S700 future mention information generation process, 800 future event information presentation screen

Claims (14)

The information processing device includes a processor and a memory.
A storage unit that stores event-related mention information, which is information related to an event-related mention that is a mention of an event included in document information;
a user mention extraction unit that extracts user mentions, which are mentions related to event designation information, which is information designating a predetermined event, from the event-related mention information;
a future mention identification unit that identifies a future mention, which is a mention regarding a future event regarding the event represented by the user mention, from the event-related mention information;
Equipped with
The user mention extraction unit extracts a plurality of the user mentions from the event-related mention information, and generates a user mention set that is a set including the extracted plurality of the user mentions as elements;
The future mention identification unit extracts a plurality of mentions that are candidates for the future mention from the event-related mention information as candidates, generates a plurality of future mention sets that are sets obtained by classifying the extracted candidates, and calculates a causal strength between the user mention set and each of the plurality of future mention sets.
Future event estimation system. 2. The future event estimation system according to claim 1,
The event-related mention information includes a peripheral event mention, which is a mention of a certain event and a peripheral event that is an event related to the certain event;
The future mention identification unit identifies, as the future mention, a mention regarding an event similar to the event represented by the user mention, from the event-related mention information, based on a similarity to the peripheral event mention.
Future event estimation system. 2. The future event estimation system according to claim 1,
the event-related mention information includes time information that is information indicating a time when the event represented by the event-related mention occurred;
the future mention identification unit identifies, as the future mention, a mention regarding an event similar to the event represented by the user mention from the event-related mention information, based on a similarity with the time information;
Future event estimation system. 2. The future event estimation system according to claim 1 ,
a future event information generation unit that outputs the contents of the future mention set having a high causal strength to a user interface with priority;
The future event estimation system further comprises: 5. The future event estimation system according to claim 4 ,
The future event information generation unit identifies a representative mention for each of the future mention sets, and outputs the identified mention to a user interface in an emphasized manner.
Future event estimation system. 6. The future event estimation system according to claim 5 ,
the event-related mention information includes time information that is information indicating a time when the event represented by the event-related mention occurred;
the future event information generation unit identifies, among the mentions belonging to the future mention set, a mention that has the earliest time identified from time information as a representative of the future mention set;
Future event estimation system. 6. The future event estimation system according to claim 5 ,
The future event information generation unit identifies, from among the mentions in the future mention set, a mention having a maximum similarity to other elements belonging to the future mention set as a representative of the future mention set.
Future event estimation system. 2. The future event estimation system according to claim 1,
a document information acquisition unit for acquiring the document information;
an event-related mention information generating unit that generates the event-related mention information by performing natural language processing on the document information;
The future event estimation system further comprises: 9. A future event estimation system according to claim 8 ,
The natural language processing is processing by at least one of predicate structure analysis and syntactic analysis.
Future event estimation system. 2. The future event estimation system according to claim 1,
an event designation information receiving unit that receives an input of the event designation information from a user via a user interface;
The future event estimation system further comprises: The future event estimation system according to claim 10 ,
an event designation information candidate extraction unit that extracts information that is a candidate for the event designation information from the event-related mention information;
an event designation information presentation unit that presents the candidate information to a user via a user interface in a selectable state when the event designation information reception unit receives an input of the event designation information;
The future event estimation system further comprises: An information processing device having a processor and a memory,
A step of storing event-related mention information, which is information regarding an event-related mention that is a mention of an event included in the document information;
A step of extracting user mentions, which are mentions related to event designation information, which is information designating a predetermined event, from the event-related mention information;
Identifying a future mention from the event-related mention information, the future mention being a mention of a future event regarding the event represented by the user mention;
extracting a plurality of the user mentions from the event-related mention information, and generating a user mention set that is a set having the extracted plurality of the user mentions as elements;
extracting a plurality of references that are candidates for the future reference from the event-related reference information as candidates, generating a plurality of future reference sets that are sets obtained by classifying the extracted candidates, and determining a causal strength between the user reference set and each of the plurality of future reference sets;
A method for estimating future events. A method for estimating a future event according to claim 12 , comprising:
The event-related mention information includes a peripheral event mention, which is a mention of a certain event and a peripheral event that is an event related to the certain event;
a step of the information processing device identifying a mention of an event similar to the event represented by the user mention as the future mention from the event-related mention information based on a similarity to the peripheral event mention;
The method for estimating future events further comprises: A method for estimating a future event according to claim 12 , comprising:
the event-related mention information includes time information that is information indicating a time when the event represented by the event-related mention occurred;
a step of identifying a mention of an event similar to the event represented by the user mention as the future mention from the event-related mention information based on a similarity with the time information by the information processing device;
The method for estimating future events further comprises:

Images