JP7628576B2 - Analysis support server, analysis support method and program - Google Patents

Description

The present invention relates to an analysis support server, an analysis support method, and a program.

Technology is known for analyzing texts written about companies in order to consider investment targets. Patent Document 1 describes a technology that calculates the corporate evaluation value of companies to be evaluated based on the evaluation values of keywords contained in evaluation texts for each of multiple companies.

JP 2007-183796 A

Information about companies, i.e., corporate information, may be collected from various data sources. If corporate information collected from various data sources is presented separately for each data source, it may be difficult for a user to efficiently analyze the corporate information. Some aspects of the present invention aim to provide a technique for assisting a user in efficiently analyzing corporate information.

According to some embodiments, an analysis support server for providing information to support the analysis of a company is provided, the analysis support server comprising: an acquisition means for acquiring a search keyword specification from a user; an identification means for identifying one or more records including the search keyword from a plurality of records read from a plurality of data sources and identifying one or more companies from the one or more records; and a display means for displaying a list showing the one or more identified companies.

Some embodiments can help users analyze corporate information more efficiently.

FIG. 1 is a block diagram illustrating an example of the configuration of a system according to some embodiments. FIG. 1 is a block diagram illustrating an example of a hardware configuration of a computer according to some embodiments. FIG. 4 is a schematic diagram illustrating various records according to some embodiments. FIG. 11 is a flow diagram illustrating an example method for generating a classification record according to some embodiments. FIG. 4 is a flow diagram illustrating an example method for searching company information according to some embodiments. FIG. 11 is a schematic diagram illustrating an example of a display screen for search results according to some embodiments.

The following embodiments are described in detail with reference to the attached drawings. Note that the following embodiments do not limit the invention according to the claims, and not all combinations of features described in the embodiments are necessarily essential to the invention. Two or more of the features described in the embodiments may be combined in any combination. In addition, the same reference numbers are used for the same or similar configurations, and duplicate descriptions are omitted.

With reference to FIG. 1, an example of the configuration of a system including an analysis support server 100 according to some embodiments will be described. The analysis support server 100 is a server that performs processing to support a user in analyzing a company. Details of this processing will be described later. The analysis support server 100 can communicate with one or more database (DB) servers through a network 107. For example, the analysis support server 100 may be able to communicate with a company DB server 101, an investor DB server 102, a literature DB server 103, an intellectual property DB server 104, and a clinical trial DB server 105. Each DB server provides information managed by itself to the analysis support server 100 in response to a request from the analysis support server 100. In this way, each DB server functions as a data source. The analysis support server 100 may use only a part of these DB servers as data sources, or may use other DB servers as additional data sources.

The company DB server 101 is a server that manages a DB that stores general information about companies. The investor DB server 102 is a server that manages a DB that stores information about investors. An investor may be a legal entity such as a company, or may be a natural person. The literature DB server 103 is a server that manages a DB that stores information about literature. For example, the literature DB server 103 may store information about technical literature. The technical literature may include academic literature such as papers. The intellectual property DB server 104 is a server that manages a DB that stores information about intellectual property. The intellectual property may include patents, designs, trademarks, etc. The clinical trial DB server 105 is a server that manages a DB that stores information about clinical trials.

The analysis support server 100 may further be capable of communicating with a client terminal 106 through a network 107. The client terminal 106 may be used by a user to utilize the analysis support service provided by the analysis support server 100. For example, the analysis support server 100 may provide the analysis support service through a web application. The client terminal 106 may access the web application of the analysis support server 100 using its own browser or a dedicated application. The client terminal 106 may be, for example, a personal computer, a smartphone, a tablet, etc.

The network 107 is any network that provides the analysis support server 100 with communication with other devices (DB servers and client terminals 106). For example, the network 107 may include a local area network (LAN), a wide area network such as the Internet, or an internal network within a cloud system.

With reference to FIG. 2, an example of the hardware configuration of the computer 200 will be described. The computer 200 may include the components shown in FIG. 2. The computer 200 may not include some of the components shown in FIG. 2, or may include components not shown in FIG. 2. The computer 200 may be used as any of the analysis support server 100, the company DB server 101, the investor DB server 102, the literature DB server 103, the intellectual property DB server 104, the clinical trial DB server 105, and the client terminal 106. The multiple servers shown in FIG. 1 may be realized by separate computers 200. Alternatively, some of the multiple servers shown in FIG. 1 may be realized by a single computer 200.

The processor 201 is a device that controls the overall operation of the computer 200. The processor 201 functions as, for example, a CPU (Central Processing Unit). The memory 202 is a device that stores programs and temporary data necessary for the operation of the computer 200. The memory 202 is composed of, for example, a RAM (Random Access Memory) or a ROM (Read Only Memory). The operation of the computer 200 may be performed, for example, by the processor 201 executing a program stored in the memory 202. Alternatively, some or all of the operation of the computer 200 may be performed by a dedicated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

The input device 203 is a device for obtaining input from a user of the computer 200. The input device 203 is, for example, a keyboard or a mouse. The output device 204 is a device for outputting to the user of the computer 200. The output device 204 is, for example, a display or a speaker. The communication device 205 is a device for connecting the computer 200 to the network 107. The communication device 205 is, for example, a network interface card (NIC) or a wireless communication module.

The storage device 206 is a device that stores data used for the operation of the computer 200. The storage device 206 is composed of a storage medium such as a hard disk drive (HDD), a solid state drive (SSD), or a digital versatile disk (DVD).

With reference to FIG. 3, the information managed by the analysis support server 100 and DB servers 101 to 105 will be described. The corporate DB server 101 manages a corporate record 301 for each company. In general, the corporate DB server 101 manages multiple corporate records 301. A corporate record 301 is a record that represents information about an individual company. The multiple corporate records 301 may be managed in a table format or in another format. The same applies to the other records described below.

Company record 301 has the items shown in FIG. 3. Company record 301 may not include some of the items shown in FIG. 3, or may include other items. The same applies to the other records described below. The description of company record 301 is a sentence that describes the company, and is written in natural language. The keywords of company record 301 are one or more words that represent the company. Information for each item of company record 301 may be registered by an administrator who manages the company DB server 101.

The investor DB server 102 manages an investor record 302 for each investor. In general, the investor DB server 102 manages multiple investor records 302. An investor record 302 is a record that represents information about an individual investor. Information for each item of the investor record 302 may be registered by an administrator who manages the investor DB server 102. The investor record 302 may include, for example, the investor name. The investor name may be the name of a natural person or legal entity (e.g., a company) making an investment.

The literature DB server 103 manages a literature record 303 for each document. In general, the literature DB server 103 manages multiple literature records 303. A literature record 303 is a record for managing information about an individual document. Information for each item of the literature record 303 may be registered by an administrator who manages the literature DB server 103. The literature record 303 may include, for example, an author name and an affiliation name. The author name may be the name of the natural person who created the document, and the affiliation name may be the name of the company to which this natural person belongs.

The intellectual property DB server 104 manages an intellectual property record 304 for each intellectual property right case. In general, the intellectual property DB server 104 manages multiple intellectual property records 304. An intellectual property record 304 is a record for managing information on individual cases related to intellectual property rights (e.g., patent application cases, design registration application cases, etc.). Information for each item of the intellectual property record 304 may be registered by an administrator who manages the intellectual property DB server 104. The intellectual property record 304 may include, for example, the name of the applicant and the name of the patent owner. The applicant may be the name of a natural person or legal entity (e.g., a company) that has filed the application. The right owner name may be the name of a natural person or legal entity (e.g., a company) that holds the intellectual property right.

The clinical trial DB server 105 manages a clinical trial record 305 for each clinical trial. In general, the clinical trial DB server 105 manages multiple clinical trial records 305. A clinical trial record 305 is a record for managing information about an individual clinical trial. Information for each item of the clinical trial record 305 may be registered by an administrator who manages the clinical trial DB server 105. The clinical trial record 305 may include, for example, the name of the implementer, the name of the sponsor, and the name of the collaborator. The name of the implementer may be the name of the company that implements the clinical trial. The name of the sponsor may be the name of the company that invests in the clinical trial. The name of the collaborator may be the name of the company that cooperates in the clinical trial.

The analysis support server 100 manages the classification record 300 for each company. In general, the analysis support server 100 manages multiple classification records 300. The classification record 300 is a record for managing the classification of an individual company. Information for each item of the classification record 300 may be registered by an administrator who manages the analysis support server 100, or may be set by the analysis support server 100. The analysis support server 100 may manage the classification records 300 for all of the multiple company records 301 managed by the company DB server 101, or may manage the classification records 300 for some of the multiple company records 301 managed by the company DB server 101. For example, the analysis support server 100 may manage the classification of companies related to the healthcare and medical fields among the multiple company records 301 managed by the company DB server 101. One classification record 300 corresponds to any one of the company records 301.

The company record ID of the classification record 300 is identification information for uniquely identifying the company record 301 managed by the company DB server 101. The company record ID may be the company name included in the company record 301, or the legal entity name included in the company record 301. Alternatively, if unique identification information is assigned to each company record 301 by the company DB server 101, the company record ID may be this identification information.

The analysis support server 100 assigns each company to one of multiple segments. In other words, the analysis support server 100 classifies multiple companies, for which information is obtained from multiple data sources (e.g., DB servers 101 to 105), into multiple segments. The analysis support server 100 may classify each company into multiple segments at a single level, or into multiple segments in a hierarchical structure with multiple levels.

In some embodiments, the analysis support server 100 may classify each company in a four-level hierarchical structure (e.g., sector, segment, tech segment, detailed segment). In this example, sector is the top level, the next level is segment, the next level is tech segment, and the lowest level is detailed segment. These four-level hierarchical structures may have a tree structure. Specifically, each sector may be divided into multiple segments, each segment may be divided into multiple tech segments, and each tech segment may be divided into multiple detailed segments.

The highest level, sector, represents a classification of companies by business model. For example, a sector may have classifications such as B2C (business-to-customer) and B2B (business-to-business).

The second level, segments, represent classifications of technology areas on which each company strategically focuses within its business model. For example, segments may have classifications such as healthcare and Internet of Things (IoT).

The third level, Tech Segment, represents a classification by technology/service within a technology area of focus. For example, a Tech Segment may have a classification such as biometric monitoring.

The detailed segment, which is the lowest level, represents a classification of component technologies within a technology unit or a specific service unit. For example, a detailed segment may have a classification such as disease monitoring.

With reference to FIG. 4, the operation of the analysis support server 100 for classifying the companies represented by the company records 301 of the company DB server and generating the classification records 300 will be described. Each step in FIG. 4 may be performed by the processor 201 of the computer 200 operating as the analysis support server 100 executing a program read into the memory 202. In this case, the processor 201 is the processing entity that executes the operation of each step. Alternatively, some or all of the steps of the method in FIG. 4 may be executed by a dedicated circuit such as an application specific integrated circuit (ASIC). In this case, the dedicated circuit is the processing entity that executes the operation of each step. The same applies to the operation in FIG. 5 below. The operation in FIG. 4 may be started in response to an instruction from the administrator of the analysis support server 100. It is assumed that a plurality of company records 301 are stored in the company DB server 101 at the start of the operation in FIG. 4.

In S401, the analysis support server 100 reads out multiple company records 301 from the corporate DB server 101 and stores them in the storage device 206 for subsequent processing. The analysis support server 100 may read out all company records 301 managed by the corporate DB server 101, or may read out only a portion of the company records 301 (for example, companies related to the healthcare and medical fields).

At S402, the analysis support server 100 generates a classification record 300 for each of the multiple companies whose company records 301 have been read, and stores the classification record 300 in the storage device 206. The analysis support server 100 sets the company name, legal entity name, or other identification information of the individual company record 301 as the company record ID of the individual classification record 300. At this point, other items in the classification record 300 (e.g., sector, etc.) are blank.

In S403, the analysis support server 100 sets a sector for each of the multiple companies for which the classification record 300 has been generated, and stores this sector in the sector item of the classification record 300. The analysis support server 100 may set the sector for each company according to rules specified by the administrator of the analysis support server 100.

In S404, the analysis support server 100 sets a segment for each of the multiple companies in each sector and stores the segment in the segment item of the classification record 300. To set the segments, the analysis support server 100 may divide the multiple companies in each sector into multiple groups by using natural language processing by machine learning (hereinafter simply referred to as "natural language processing").

An example of a method for dividing multiple companies in each sector into multiple groups using natural language processing will be described. For example, the analysis support server 100 may divide multiple companies in each sector into multiple groups by performing text-based natural language processing. For example, the analysis support server 100 may perform natural language processing using text (e.g., descriptions) included in the company records 301 of multiple companies in each sector. Such text-based natural language processing may be performed using existing technology, for example, using top2vec. With top2vec, the context of each paragraph included in the text and the connections between the paragraphs can be interpreted.

An example of text-based natural language processing will be described below. First, the analysis support server 100 extracts descriptions from the company records 301 of multiple companies included in one sector, and generates a vector representing the description of each company record 301. This vector may be, for example, a 300-dimensional vector. Then, the analysis support server 100 converts each vector into a low-dimensional (for example, five-dimensional) vector. Finally, the analysis support server 100 divides the multiple companies included in one sector into multiple groups by clustering the low-dimensional vectors representing the multiple companies. For example, the analysis support server 100 may score the similarity of the context and connections of the paragraphs included in the description of each company record 301, and classify the descriptions into the same group if the scores are higher than a predetermined threshold. The analysis support server 100 performs the above-mentioned processing for each of the multiple sectors.

Then, the analysis support server 100 sets a segment name for each group. The segment name may be determined by the analysis support server 100 according to a predetermined rule, or may be set by an administrator of the analysis support server 100.

For some sectors (e.g., sectors with 500 or fewer companies), the analysis support server 100 may set segments according to rules specified by an administrator of the analysis support server 100, instead of performing natural language processing as described above.

In S405, the analysis support server 100 sets a tech segment for each of the multiple companies in each segment and stores the tech segment in the tech segment item of the classification record 300. In order to set the tech segments, the analysis support server 100 may divide the multiple companies in each segment into multiple groups by using natural language processing. The division in S405 may use top2vec, as in the division in S404.

In S406, the analysis support server 100 sets detailed segments for each of the multiple companies in each tech segment, and stores the detailed segments in the detailed segment items of the classification record 300. For example, the analysis support server 100 may perform natural language processing using words included in the company records 301 of the multiple companies in each tech segment. The words used in the natural language processing may be words extracted from the descriptions of the company records 301, or may be keywords in the company records 301. Such word-based natural language processing may be performed using existing technology, for example, using word2vec.

An example of word-based natural language processing will be described below. First, the analysis support server 100 extracts descriptions from the company records 301 of multiple companies included in one tech segment. The analysis support server 100 generates a model (word2vec model) by performing machine learning on the words included in the descriptions. This learning may be learning using only data acquired from the corporate DB server 101, or additional learning for an already trained model using a general-purpose corpus.

Then, the analysis support server 100 obtains a vector by inputting the keywords included in the company record 301 into the model. This vector may be, for example, a 300-dimensional vector. If the company record 301 contains multiple keywords, the analysis support server 100 may generate a vector based on the vectors of each of these keywords (for example, by extracting the maximum value or averaging).

Then, the analysis support server 100 converts each vector into a low-dimensional (e.g., five-dimensional) vector. Finally, the analysis support server 100 divides the multiple companies included in one tech segment into multiple groups by clustering the low-dimensional vectors representing the multiple companies. This clustering may be performed, for example, by the k-means method. The number of clusters may be set by the administrator of the analysis support server 100. The analysis support server 100 performs the above-mentioned process for each of the multiple tech segments.

Then, the analysis support server 100 sets a detailed segment name for each cluster. The detailed segment name may be determined by the analysis support server 100 according to a predetermined rule, or may be set by an administrator of the analysis support server 100.

As described above, the analysis support server 100 classifies segments and tech segments and detailed segments using different methods. Specifically, the upper hierarchical segments and tech segments are classified using a sentence-based method, and the lower hierarchical segments are classified using a word-based method. The lower the hierarchical level, the fewer the number of companies to be classified. Through experiments, the inventors found that the sentence-based method can classify more accurately when there are many objects to be classified, and that the word-based method can classify more accurately when there are few objects to be classified. Therefore, by performing classification using different methods as described above, accurate classification can be achieved.

With reference to FIG. 5, an example of the operation of the analysis support server 100 to provide information on companies that satisfy certain conditions will be described. The operation of FIG. 5 may be started in response to an instruction from a user of the analysis support server 100. At the start of the operation of FIG. 5, it is assumed that each item of FIG. 3 is stored in the DB servers 101 to 105 and the analysis support server 100.

In S501, the analysis support server 100 generates a screen for acquiring search conditions and presents it to the user. For example, when a user is using the analysis support server 100 via a client terminal 106, the analysis support server 100 may transmit such a screen to the client terminal 106, which may then display the screen on its own output device 204. Alternatively, when a user is using the analysis support server 100 directly, the analysis support server 100 may display such a screen on its own output device 204.

An example of a screen 600 for acquiring search conditions will be described with reference to FIG. 6. Screen 600 may include, for example, a pull-down list 601 for acquiring a designation of a database to be searched, a text box 602 for acquiring search keywords, a button 603 for acquiring an instruction to execute a search, and a display area 604 for displaying search results. At the time of S601 (i.e., before the search is executed), nothing may be displayed in display area 604.

In S502, the analysis support server 100 acquires the designation of the database to be searched and the search keyword from the user through the pull-down list 601 and text box 602 on the screen 600. Instead of the pull-down list 601, another graphical object for accepting the selection of one item from multiple items, for example a list with check boxes, may be used. The database to be searched may be selectable from five options, for example, "company information," "investor information," "literature information," "intellectual property information," and "clinical trial information." The search keyword may be any word. For example, the search keyword may be a company name or a disease name.

In S503, the analysis support server 100 executes a search using the designation of the database to be searched and the search keyword. For example, assume that "company information" is designated as the database to be searched. In this case, the analysis support server 100 may identify a company record 301 that contains the search keyword in any of the company name, legal entity name, description, and keywords, among the multiple company records 301 managed by the company DB server 101. Alternatively, the analysis support server 100 may identify a company record 301 in which the search keyword is included in other items. The analysis support server 100 then identifies, as a matching record, a record related to the company indicated by the company name of the identified company record 301, among the multiple classification records 300 managed by the analysis support server 100.

Suppose "Investor Information" is specified as the database to be searched. In this case, the analysis support server 100 may identify, from among the multiple investor records 302 managed by the investor DB server 102, investor records 302 in which the search keyword is included in either the investor name or keywords. Alternatively, the analysis support server 100 may identify investor records 302 in which the search keyword is included in other items. The analysis support server 100 then identifies, from among the multiple classification records 300 managed by the analysis support server 100, records related to a company indicated by the investor name of the identified investor record 302 as a matching record. Instead of or in addition to the investor name, other items representing a company may be used from among the items included in the investor record 302.

Suppose "literature information" is specified as the database to be searched. In this case, the analysis support server 100 may identify a literature record 303 in which the search keyword is included in any of the journal name, abstract, or keywords, among the multiple literature records 303 managed by the literature DB server 103. Alternatively, the analysis support server 100 may identify a literature record 303 in which the search keyword is included in other items. The analysis support server 100 then identifies, as a matching record, a record related to the company indicated by the affiliation name of the identified literature record 303, among the multiple classification records 300 managed by the analysis support server 100. Instead of or in addition to the affiliation name, other items representing companies may be used among the items included in the literature record 303.

Suppose that "intellectual property information" is specified as the database to be searched. In this case, the analysis support server 100 may identify, from among the multiple intellectual property records 304 managed by the intellectual property DB server 104, intellectual property records 304 in which the search keyword is included in any of the name, abstract, claims, and keywords. Alternatively, the analysis support server 100 may identify intellectual property records 304 in which the search keyword is included in other items. The analysis support server 100 then identifies, from among the multiple classification records 300 managed by the analysis support server 100, records related to a company indicated by the applicant name or right holder name of the identified intellectual property record 304 as matching records. Instead of or in addition to the applicant name or right holder name, other items representing a company may be used from among the items included in the intellectual property record 304.

Suppose that "clinical trial information" is specified as the database to be searched. In this case, the analysis support server 100 may identify a clinical trial record 305 in which the search keyword is included in any of the sponsor name, collaborator name, and keyword among the multiple clinical trial records 305 managed by the clinical trial DB server 105. Alternatively, the analysis support server 100 may identify a clinical trial record 305 in which the search keyword is included in other items. Then, the analysis support server 100 identifies, among the multiple classification records 300 managed by the analysis support server 100, a record related to a company indicated by the implementer name, sponsor name, or collaborator name of the identified clinical trial record 305 as a matching record. Instead of or in addition to the implementer name, sponsor name, or collaborator name, other items representing a company among the items included in the clinical trial record 305 may be used.

In S504, the analysis support server 100 displays the search results in the display area 604 of the screen 600. For example, as shown in FIG. 6, the analysis support server 100 includes in the display area 604 a message 605 showing an overview of the search results and a table 606 of hit records. Each record in the table 606 represents a matching record identified in S503. The table 606 includes multiple columns 607-612. Column 607 indicates the database specified in the pull-down list 601. Column 608 indicates the sector of the identified company. Column 609 represents a segment of the identified company. Column 610 represents the tech segment of the identified company. Column 611 represents a detailed segment of the identified company. Column 612 represents the company name of the identified company.

Each field of columns 610, 611, and 612 may be a link to a detailed page. In response to a user selecting one of these links, the analysis support server 100 may present the user with a page showing details of the tech segment, detailed segment, or company. The records of table 606 may be sortable using each of columns 608 to 612 as a key. The records of table 606 may be filtered for each of columns 609 to 611.

As described above, according to screen 600 in FIG. 6, records containing the search keywords are extracted from multiple data sources, and a list of companies associated with these records is displayed. In addition, the sectors, segments, tech segments, and detailed segments are displayed along with the company names of companies that meet the search criteria. Furthermore, data sources that store records containing the search keywords are displayed along with the company names of companies that meet the search criteria. This allows the user to accurately grasp the trends of companies that meet the specified search criteria.

The invention is not limited to the above-described embodiment, and various modifications and variations are possible within the scope of the invention.

100 Analysis support server, 101 Corporate DB server, 106 Client terminal

Claims (4)

An analysis support server for providing information for supporting analysis of a company,
An acquisition means for acquiring a search keyword specification from a user;
an identification means for identifying one or more records including the search keyword from a plurality of records read from a plurality of data sources, and identifying one or more companies from the one or more records;
A classification means for classifying a plurality of companies into each of a plurality of hierarchies including a first hierarchical level and a second hierarchical level lower than the first hierarchical level;
a display means for displaying a list showing the one or more identified companies ;
The classification means classifies companies in the first layer using sentence-based natural language processing, and classifies companies in the second layer using word-based natural language processing;
The display means displays, for each of the plurality of hierarchical levels, classification items assigned to a company shown in the list in association with the company . The analysis support server according to claim 1, wherein the display means displays the data source from which the record used to identify the company shown in the list was read in association with the company. A program for causing a computer to function as each of the means of the analysis support server according to claim 1 or 2 . 1. A computer-implemented method for providing information to support an analysis of a company, the method comprising:
An acquisition step of acquiring a search keyword specification from a user;
an identifying step of identifying one or more records including the search keyword from a plurality of records read from a plurality of data sources, and identifying one or more companies from the one or more records;
A classification step of classifying a plurality of companies into each of a plurality of hierarchies including a first hierarchical level and a second hierarchical level lower than the first hierarchical level;
and displaying a list showing the one or more identified companies ,
The classification step includes classifying the companies in the first hierarchical level using sentence-based natural language processing and classifying the companies in the second hierarchical level using word-based natural language processing;
The analysis support method , wherein the display step includes displaying, for each of the plurality of hierarchical levels, classification items assigned to a company shown in the list in association with the company .