JP4837989B2 - Document information analysis apparatus and document information analysis method - Google Patents

Description

  The present invention relates to a document information analysis apparatus and a document information analysis method for analyzing document information such as patent documents and academic documents, for example.

  For example, by investigating patent documents as document information, it is possible to find out the direction of research and development or to make business plans. For this purpose, patent documents that match a predetermined purpose are extracted from a large group of patent documents, and the extracted patent documents are analyzed. A representation of this analysis result on a two-dimensional plane is known as a so-called patent map, and a technique for automatically creating a patent map from a group of patent documents is known (Patent Document 1).

In addition, there is also known a technique that can analyze time-series changes in the number of published patent documents (Patent Document 2).
JP 2005-149346 A JP 2001-092825 A

  In the technique described in the first document, a patent map in which keywords are arranged at key points can be automatically created based on a large number of patent documents that match a search condition designated by the user. Since there is no means to predict, there is room for improvement in usability.

  According to the technique described in the second document, in the matrix-type patent map, only the increase / decrease tendency of the application for each cell is analyzed, and the matrix structure does not change. It cannot be visualized and analyzed and is not easy to use.

  As described above, in any of the conventional technologies, it is not possible to provide a determination material that indicates in what direction and at what speed, in the entire technical field to which the technology focused on by the user belongs, Not suitable for future predictions.

  Accordingly, an object of the present invention is to provide a literature information analysis apparatus and literature information that can predict a temporal change of a map based on a past map and a prediction model and visualize and express a future map change. It is to provide an analysis method. Another object of the present invention is to predict a future map using a prediction model selected from a plurality of prediction models prepared in advance, and visually express a temporal change of the map from the viewpoint of the course and the timing. An object of the present invention is to provide a literature information analysis apparatus and a literature information analysis method capable of performing the above. The further objective of this invention will become clear from description of embodiment mentioned later.

  In order to solve the above problems, a literature information analysis apparatus according to one aspect of the present invention predicts a time change of a map generated by a map control unit that generates a predetermined map based on map generation conditions, and a map control unit And a predicting unit. The map control unit includes a document information storage unit that stores a plurality of document information that has been digitized, and a predetermined plurality of document information that is extracted from a predetermined document information group determined by map generation conditions among the stored document information. A document coordinate calculation unit that calculates the coordinates of each document information of a predetermined document information group by performing principal component analysis based on the combination of keywords and the number of occurrences, the total number of document information including each predetermined keyword, and By performing principal component analysis based on the number of occurrences of each predetermined keyword, a keyword coordinate calculation unit that calculates the coordinates of each predetermined keyword, and based on the coordinates of each document information calculated by the document coordinate calculation unit Calculates the distribution density of each document information, and visualizes map figures and contours with contours based on the calculated distribution density. And a, a map generation unit which generates map information Te. Based on the first map graphic generated by the map control unit based on a predetermined document information group and the set prediction model, the prediction unit determines the second map graphic at a predetermined time point in the future than the first map graphic. A prediction processing unit that performs prediction. Then, the map information is combined with the guide information for guiding the change from the first map graphic to the second map graphic, and the route from the first map graphic to the second map graphic is displayed in this guide information. The information of at least is included.

  Examples of the document information include patent documents (including open gazettes, registered gazettes) and academic papers. The document information is digitized and stored in the document information storage unit. Among a plurality of document information stored in the document information storage unit, a plurality of predetermined keywords are extracted from a predetermined document information group. The predetermined document information group can be defined as a group of document information that matches a map generation condition (search condition) specified by the user, for example. The predetermined plural keywords are, for example, main keywords that characterize each piece of document information in a predetermined document information group.

  The document coordinate calculation unit performs principal component analysis based on a combination of a plurality of predetermined keywords and the number of appearances of the plurality of predetermined keywords, and calculates the coordinates of each document information in a predetermined document information group. Principal component analysis is a method of multivariate analysis. To put it simply, each sample information that contains many variables is represented by several comprehensive indicators (principal components) that most directly represent the differences. This is an analysis method that reduces the number of dimensions.

  Similarly, the keyword coordinate calculation unit calculates the coordinates of each predetermined keyword by principal component analysis. The map generation unit calculates the distribution density of the document information based on the coordinates of each document information. For example, the map generation unit can detect the density of the document information for each block by dividing the entire area of the map into a plurality of blocks and calculating the number of document information located in each block. For example, the map generation unit creates a map figure having an outline by associating display elements (for example, contour lines) with the distribution density of the document information. The map generation unit generates map information by arranging the map graphic and a display element indicating the presence of the keyword (for example, a keyword character itself or a symbol) on the map area. This map information is expressed, for example, on a two-dimensional plane, but is not limited thereto, and can be expressed in a three-dimensional space.

  The prediction unit predicts the second map graphic based on the first map graphic and the prediction model. The first map graphic is a map graphic generated based on a predetermined document information group, which is obtained from document information already accumulated from the past to the present, and may be called a fixed map graphic. it can. The second map graphic is a predicted map graphic and can be called an undefined map graphic.

Here, the point to be noted is that the second map graphic is also arranged so as to overlap the coordinate system in which the first map graphic is arranged. That is, the second map graphic is also arranged in a coordinate system based on a predetermined document information group that is the basis of the first map graphic. Major keywords are also displayed in the map information. Therefore, the user can change from the first map graphic to the second map graphic in the technical field according to the positional relationship between the first map graphic and the second map graphic existing in the same coordinate system and the keywords displayed in the vicinity thereof. Changes can be observed and judged.
The first map graphic and the second map graphic are formed so as to connect blocks having a distribution density equal to or higher than a predetermined threshold. That is, the first map graphic and the second map graphic are generated as main map graphics, and other document information and keywords exist outside the first map graphic and the second map graphic in the map space. ing.

  Next, it should be noted that the map information includes guidance information, and the guidance information includes route information from the first map graphic to the second map graphic. The course information can be defined as information indicating the direction and distance from the first map graphic to the second map graphic, for example. The course information can be positively displayed as a display element (for example, an arrow) different from the first and second map figures, for example. Alternatively, the route information may be embedded in the change status of each map graphic. For example, during the period from the first map figure as the starting point (reference) to the second map figure in the final state, a plurality of other map figures showing intermediate transition states are displayed continuously or intermittently. By doing so, the course from the first map graphic to the second map graphic and the distance between the first and second maps can be expressed.

  The guide information may further include information indicating the time when the first map graphic changes to the second map graphic. For example, the time to reach each stage (each intermediate map figure) from the first map figure as a reference to the second map figure in the final state can be displayed. Thereby, the user can grasp the time when the change occurs, for example, the time when a new technical element is taken in (expansion time), the time when an existing technical element is abandoned (withdrawal time), and the like.

  The prediction unit may include a prediction model setting unit that selects and sets at least one prediction model specified by the prediction condition from among a plurality of prediction models prepared in advance. As the prediction model, for example, both a linear prediction model and a nonlinear prediction model can be prepared in advance, and at least one of the prediction models can be selected and set. Examples of the linear prediction model include a least square method. Examples of the nonlinear prediction model include a movement model based on the analysis result of a similar technique.

  The predicting unit predicts a second map graphic at each of a plurality of future points in time, and each second map graphic is visualized on the coordinate space where the first map graphic is displayed together with the first map graphic. Arranged. For example, when the user instructs prediction after 10 years, the prediction unit predicts the second map graphic every year from the present time, and arranges each second map graphic in the map information. Here, if the second map graphic in the final state is called the final prediction map graphic, and the first map graphic serving as a reference reflecting the latest document information group is referred to as the reference map graphic, the final prediction map is converted from the reference map graphic. An intermediate prediction map graphic can be generated and arranged at each stage up to the graphic (for each predetermined period). In a normal case, these intermediate prediction map figures are arranged in a map space (a coordinate system obtained from a predetermined document information group that is the basis of the first map figure) so as to overlap each other. Therefore, the user can grasp the state of continuous change from the reference map graphic to the final prediction map graphic by the intermediate prediction map graphic continuously arranged so as to overlap each other.

  The prediction unit divides a predetermined document information group at predetermined intervals, and generates a first map graphic based on the document information group for each predetermined period by the map control unit. It is also possible to generate map information by visualizing it with a map figure. In the above example, by generating multiple intermediate past map figures that go back to the past from the reference map figure, it is possible to visually see how the map figure changes continuously from the past to the future. Can be expressed.

  The prediction unit divides a predetermined document information group every predetermined period, and generates a first map graphic based on the document information group for each predetermined period by the map control unit, and the time change of each first map graphic And at least one second map figure can be predicted based on the set prediction model. That is, in the above example, the prediction unit generates the intermediate prediction map graphic or the final prediction map graphic based on the temporal change from the intermediate past map graphic to the reference map graphic and the prediction model. Can do.

  The prediction unit determines at least the position and shape of the second map graphic based on the locus of the center of gravity of each first map graphic for each predetermined period, the shape change of each first map graphic, and the set prediction model. One or more can be predicted. In the above example, the position and shape of the intermediate predicted map graphic or the final predicted map graphic can be predicted based on the locus of the center of gravity position and the shape change of the reference map graphic and the intermediate past map graphic. For example, when it is considered that the shape does not change greatly, a configuration in which the shape of the reference map figure that is the latest determined shape is used as it is may be used.

  The prediction unit causes the map control unit to create a predetermined map figure based on the document information group created by the selected predetermined document creator among the document information group stored in the document information storage unit, It is also possible to compare the predetermined map graphic with the first map graphic and the second map graphic and output the comparison result. For example, when the user designates a document creator such as a specific company or university, the prediction unit generates a predetermined map graphic based on the document information group created by the document creator. This predetermined map figure can also be called a specific organization map, for example. By comparing the specific organization map with the first map graphic and the second map graphic, the user can grasp, for example, what position a predetermined document creator occupies in the technical field. Can do.

  The map control unit can also calculate the coordinates of each piece of document information in the predetermined document information group by weighting the document information group having a higher usage degree in the predetermined document information group so that the score becomes higher. . The document information group having a high degree of use means, for example, a document information group having a high use frequency or a document information group having a high use value. Specifically, document information frequently cited in other documents, document information having a large number of requests for browsing, and the like are included. Since such highly utilized document information or document information group is weighted according to the level of utilization, the map space can be set reflecting its value.

  In the embodiment of the present invention, a difference detection unit that detects a difference between a plurality of map graphics created by the map control unit is provided. For example, the difference detection unit calculates a difference between a plurality of map graphics, such as a difference between a first map graphic and a second map graphic, a difference between each intermediate predicted map graphic, a difference between each intermediate past map graphic, and the like. To detect. This detected difference can be displayed in the map information.

  According to another aspect of the present invention, a computer program includes a map control unit that generates a predetermined map based on a specified map generation condition, and a prediction unit that predicts a time change of a map generated by the map control unit. To function as. Then, the map control means is a predetermined document information extracted from a predetermined document information group determined by the map generation condition among the respective document information stored in the document information storage means for storing each of the plurality of digitized document information. Document coordinate calculation means for calculating the coordinates of each document information of a predetermined document information group by performing principal component analysis based on a combination of a plurality of keywords and the number of appearances, and the total number of document information including each predetermined keyword And by performing principal component analysis based on the number of occurrences of each predetermined keyword, the keyword coordinate calculating means for calculating the coordinates of each predetermined keyword, and the coordinates of each document information calculated by the document coordinate calculating means The distribution density of each document information is calculated, and a map figure having an outline based on the calculated distribution density and a predetermined keyword It includes a map generating means for generating map information by visualizing respectively, as each internal function. The prediction means includes a prediction model setting means for selecting and setting any one prediction model from a plurality of prediction models prepared in advance based on the designated prediction condition, and a predetermined document by the map control means. Prediction processing means for predicting a second map graphic at a predetermined time in the future from the first map graphic based on the first map graphic generated based on the information group and the set prediction model, respectively. As prepared. The map information includes information indicating the course and time from the first map graphic to the second map graphic.

  A literature information analysis method according to still another aspect of the present invention includes a step of accepting a prediction condition including a prediction model, a prediction target period, and at least one keyword, and a literature information group held so far. A step of extracting a predetermined document information group including a keyword in the prediction condition, and a predetermined period belonging to each predetermined period for every predetermined period from a predetermined time point in the past to the present time indicated in the prediction target period. Based on the document information group, a second map at a predetermined time point in the future from the present time is generated based on the step of generating the first map graphic, the prediction model specified by the prediction condition, and each generated first map graphic. Generating a map graphic, and displaying each first map graphic and the second map graphic on the same screen. Then, the step of generating the first map graphic includes performing a principal component analysis based on a combination of a plurality of predetermined keywords included in the extracted predetermined document information group and the number of appearances, thereby obtaining a predetermined document information group. A step of calculating the coordinates of each document information, a step of calculating the distribution density of each document information based on the calculated coordinates of each document information, and generating a map figure based on the calculated document density The step of generating the second map graphic includes calculating the centroid position of each first map graphic and inputting the locus of each centroid position to the prediction model, thereby generating the second map graphic. It is designed to generate.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. The literature information analysis apparatus 1 according to the present embodiment can be configured as a computer apparatus including a map control unit 10, a prediction unit 20, and an information storage unit 30. The document information analysis apparatus 1 is connected to the client terminal 2 so as to be capable of bidirectional communication via a communication network such as the Internet or a LAN (Local Area Network). The client terminal 2 can be configured as, for example, a personal computer or a portable information terminal (including a mobile phone).

  The map control unit 10 generates and outputs a technical map 100 based on a large number of document information stored in the information storage unit 30. The generated map 100 is transmitted to the client terminal 2 via the communication network. The map 100 may be stored in the client terminal 2 or may not be stored in the client terminal 2. Details of the map control unit 10 will be described later.

  The prediction unit 20 predicts future changes in the map 100 based on prediction conditions designated by the user. The prediction method will be further described later.

  The information storage unit 30 stores a large number of document information such as patent publications, registration bulletins, and scientific and technical papers.

  The configuration of the map 100 will be described. The map 100 is generated according to the purpose instructed by the user. For example, the user can select only a predetermined document information group from among a large number of document information groups stored in the information storage unit 30 by designating a technical field desired to be investigated, a specific search word, or the like. it can. The content of the document information group selected by the user is analyzed by the map control unit 10 and a map 100 is generated.

  The map 100 includes one or more map figures M. The map graphic M is a graphic element enclosed and closed by the contour lines 101A and the like, and is a generic name for the reference map graphic M0, the past map graphic M-n, and the predicted map graphic M + n.

  Although details will be described later, the reference map figure M0 reflects the latest document information group that matches the prediction condition (search condition) input by the user and the past document information group before that. The past map figure M-n shows a state before the reference map figure M0. The reference map figure M0 and the past map figure M-n correspond to the “first map figure” and are respectively generated based on the already-stored literature information group. On the other hand, the prediction map M + n indicates a state ahead of the reference map graphic M0 and corresponds to the “second map graphic”.

  Each map figure M0, M-n, M + n can include, for example, one or a plurality of contour lines 101 (a generic name of 101A, 101B, 101C), a keyword 102, and a time display 103. However, all the map figures M0, M-n, and M + n do not have to have the same structure, and may have a configuration corresponding to each property.

  For example, the past map figure M-n and the predicted map figure M + n only need to include at least the outermost contour line 101A (contour line), and the inner contour lines 101B and 101C are not necessarily required. Further, the past map figure M-n and the predicted map figure M + n do not need to include the keyword 102 for characterizing itself. In addition to the main keywords 102, other potential keywords are also arranged in the map 100 as necessary. The user can easily know the positions of the past map figure M-n and the predicted map figure M + n based on the potential keywords.

  The time display 103 is time information indicated by each map figure M0, M-n, M + n. The display unit of the time may be arbitrarily set by the user, or from a predetermined unit (for example, 3 years, 1 year, 6 months, 3 months, etc.) prepared in advance in the document information analysis apparatus 1 The user may be able to select freely.

  The contour lines 101A, 101B, and 101C are display elements indicating the distribution density of the document information included in the map 100. For example, the contour line 101A indicates that the number of document information existing therein is 1 or more and less than B1, and the contour line 101B indicates that the number of document information existing therein is B1 or more and less than B2 (B1, B2). Is a natural number). As described above, the number of document information included therein increases step by step from 101A to 101B and 101C.

  A plurality of keywords 102 are also displayed on the map 100. These keywords 102 are main keywords that characterize each piece of document information of the document information group constituting the map 100, and may be referred to as the main keywords 102 in the following description.

  As described above, the map 100 displays three types of map figures, that is, the reference map figure M0, the past map figure Mn, and the predicted map figure M + n. These map figures M0, M-n, and M + n are connected by a route display 104. This course display 104 shows the relationship between the map figures M0, M-n, and M + n, and indicates the direction (path) of the map figure over time. The time display 103 and the course display 104 correspond to “guidance information”.

  Therefore, the user can confirm the past map figure M-n indicating the past state and the predicted map figure M + n indicating the future state in addition to the reference map figure M0 indicating the current state. Furthermore, the time display 103 and the course display 104 allow the user to easily understand how and in which direction the technology trend changes. Therefore, the user can make future research and development policies and business plans based on the map 100. Hereinafter, more detailed examples of the present embodiment will be described. It should be noted that the content disclosed in Japanese Patent Application Laid-Open No. 2005-149346 can be used within the scope necessary for carrying out the present invention.

  FIG. 2 is a block diagram illustrating a functional configuration of the document information analysis apparatus 1. The map control unit 10 includes, for example, a map generation condition input reception unit 11, a document extraction unit 12, a document coordinate calculation unit 13, a keyword extraction unit 14, a keyword coordinate calculation unit 15, a map generation unit 16, and a map display. It can comprise including the part 17. FIG.

  In addition, as an example of the information storage unit 30, a document database (in the figure, database is abbreviated as “DB”) 31, a word database 32, and an index database 33 can be provided.

  The document database 31 stores a plurality of document information. The word database 32 is used as a dictionary. The index database 33 is for managing what keywords are included in each document.

  The map generation condition input receiving unit 11 is for receiving an input of a map generation condition designated by the user. Accepting means, for example, receiving information indicating map generation conditions designated by the user as electronic information via a communication interface and storing the information in a memory or the like. The map generation condition input receiving unit 11 also receives map generation conditions from a prediction processing unit 23 described later.

  The document extraction unit 12 extracts document information within a predetermined range by searching the document database 31 and the index database 33 based on conditions specified by the user or input from the prediction processing unit 23. is there. The document coordinate calculation unit 13 calculates coordinates on the map 100 by analyzing the extracted document information. For example, the document coordinate calculation unit 13 calculates the coordinates of each document on the two-dimensional plane by performing principal component analysis on the keyword combination and the number of appearances included in each extracted document information.

  The keyword extraction unit 14 extracts a plurality of keywords included therein by analyzing the extracted document information. The keyword coordinate calculation unit 15 calculates the coordinates of the extracted keywords. For example, the keyword coordinate calculation unit 15 calculates the coordinates of each keyword on the two-dimensional plane by performing principal component analysis on the total number and the total number of appearances of each piece of document information including the keyword. Principal component analysis is a technique for multivariate analysis, and is a known technique, so details are omitted.

  The map generation unit 16 generates the map 100 based on the calculated coordinates of each document information and the coordinates of each keyword. For example, the map generation unit 16 sets a large number of block areas by finely dividing a finite two-dimensional plane vertically and horizontally, and calculates the number of document information existing in each block area. Thereby, the map generation unit 16 obtains the distribution density of the document information, sets the contour lines 101A and the like corresponding to the distribution density, and creates the map figure M.

  The map display unit 17 generates a map 100 that can be visually recognized by the user by allocating predetermined display elements according to the configuration of the generated map 100, and provides the map 100 to the client terminal 2. Examples of the predetermined display element include contour lines for indicating the contour lines 101A and the like, characters for indicating each keyword, a course display, a time display, and the like. In addition, a menu display unit for performing operations on the map 100 is also added.

  The prediction unit 20 includes, for example, a prediction condition input reception unit 21, a prediction model setting unit 22, a prediction processing unit 23, and a basic data storage unit 24.

  The prediction condition input receiving unit 21 electronically receives a prediction condition designated by the user. Examples of the prediction condition include a prediction target period, a prediction model specification, and an output form specification. The prediction target period includes a period that is a basis of the prediction calculation (a calculation basic period, corresponding to “target period P12” in FIG. 7) and a future period to be predicted (a prediction period, which is shown in FIG. 7). "Corresponding to" forecast period P14 ").

  Specifically, for example, the user designates the prediction target period in a format in which the start period and the end period are specified, such as “from 1995 to 2010” or “1995 to 15 years”. The document information analysis apparatus 1 can grasp the current time information using a built-in calendar, an external time server, or the like. Therefore, the calculation basic period and the prediction period can be detected by comparing the prediction target period specified by the user with the current time. The period before the current time is the calculation basic period, and the period before the current time is the prediction period.

  The prediction model setting unit 22 selects and sets any one or a plurality of prediction models from a plurality of prediction models prepared in advance. Here, examples of the prediction model include a linear prediction model and a nonlinear prediction model. An example of the linear prediction model is a least square method. Examples of the nonlinear prediction model include a weather forecast model and a movement model based on analysis of similar technologies.

In brief, the least square method determines the parameters of the model function so that the sum of squares of the difference between the theoretical value obtained from a preset model function and the predicted centroid position is minimized. Say the way. That is, the least square method is a method of calculating the future center-of-gravity position from the center-of-gravity locus of the past map figure M-n.
A weather forecast model is a prediction model that predicts future values by applying past time series values (such as the center of gravity trajectory) to a prediction model related to state changes of linear (eg, Kalman filter) and non-linear (eg, neural network). This is a method of predicting further by using the obtained value. In the case of a weather forecast model, it is also possible to use an ensemble prediction in which a large number of prediction results are calculated by changing initial values little by little, and an ensemble value (for example, an average value) of these many prediction results is employed. By using ensemble prediction, prediction accuracy can be obtained from the fluctuation width of each prediction result, and usability is improved.

  The movement model based on the analysis of the similar technology is a method of predicting the future of the prediction target technology with reference to a past technology variation pattern similar to the prediction target technology. For example, when predicting the technical trend of an MD (Mini-Disk) player, the time series analysis result of a cassette tape recorder is referred to.

  The user can designate in advance one or a plurality of prediction models from among the plurality of prediction models as described above. If the user does not specify, the prediction model set to the initial value is automatically selected.

  The prediction processing unit 23 predicts the trend of the technology specified by the user based on the set prediction model, the past map figure M-n, and the reference map figure M0. The prediction processing unit 23 instructs the map control unit 10 to generate at least one past map figure M-n before the reference map figure M0. Then, the prediction processing unit 23 calculates the locus of the center of gravity position from the generated past map figure M-n and the reference map figure M0, applies this locus to the prediction model, and applies the prediction map figure M + n to the map control unit 10. Generate. The basic data storage unit 24 stores basic data, a prediction model, and the like used for the prediction processing unit 23 to perform prediction processing.

  As shown in the lower side of FIG. 2, the prediction result (M + n) by the prediction unit 20 is displayed on the map 100. The map 100 also displays a past map figure M-n used for generating the predicted map figure M + n. Therefore, in the map 100 shown in FIG. 2, in addition to the reference map figure M0, a plurality of past map figures M-1, M-2 and a plurality of predicted map figures M + 1, M + 2 are displayed in an overlapping manner.

  The predicted map figure M + 1 and the like and the past map figure M-1 and the like are arranged in the same coordinate system as the reference map figure M0. In many cases, since the technical trend gradually changes, the map figures M + 2, M + 1, M0, M-1, and M-2 are arranged such that at least some of them overlap each other. Therefore, the map 100 shows how the trend of the prediction target technology continuously changes from the past to the future.

  Since each map figure is arranged so as to overlap, there is no need for a special route display indicating the transition direction of the map figure such as an arrow line. This is because the overall shape of the overlapping map figures also serves as a route display. However, for example, a line connecting the centroid positions of the map figures may be displayed on the map 100, and the lines connecting the centroid positions may be used as the course display.

  Although omitted in FIG. 2 due to space limitations, as shown in FIG. 1, the time display indicating the time of each map graphic is also displayed on the map 100. The time display indicates the creation time (or publication time) of the document information group from which the map graphic is derived. For example, when created based on literature information created or released from 1995 to 2005, a time display such as “1995-2005” is attached to the map graphic. The time display and the course display may be positively displayed on the map 100, or may be hidden in a normal case and displayed only when instructed by the user.

The map 100 includes markers 110A, 110B, 110C, and 110D indicating the position of a specific company in addition to the map figures M-2, M-1, M0, M + 1, and M + 2 (hereinafter referred to as the marker 110 unless otherwise distinguished). Is also displayed.
Here, the outlines of the map figures M-2, M-1, M0, M + 1, and M + 2 do not indicate the entire area in the generation year of the map graphic, but indicate the main areas in the generation year. . Therefore, other document information and keywords are scattered outside the outline of each map figure M-2, M-1, M0, M + 1, M + 2. Document information and keywords that are not included in the main map figures M-2, M-1, M0, M + 1, and M + 2 can be called non-main document information and non-main keywords, for example.
However, even if it is non-major, it is not major in the sense that it is not included in the map graphic. From a different perspective, it can be considered that the document information or keyword is ahead of the times, or the document information or keyword is outdated.
The case where the map figure M (2005) is generated based on the 2005 literature information group will be specifically described as an example. Assuming that the total number of document information created in the generation year of this map figure M (2005) is BN1 (for example, 10,000), this BN1 document information is analyzed and its distribution density is calculated. Then, for example, the map figure M (2005) is generated so as to draw the outline of a block group in which document information of threshold Thn (for example, 30) or more exists. This map figure M (2005) shows a high-density area and main areas including main keywords included in the literature information created in the year. However, the outside of the map figure M (2005) is not in a vacuum state, and there are many blocks of less than 30 cases, and non-major document information and keywords exist in these blocks.

  The marker 110 indicates the position of a specific document information group created or disclosed by a specific company, a researcher, or the like (document creator) among predetermined document information groups serving as the basis of the map 100. That is, the average value of the coordinates of each document information of a specific document information group. Therefore, the user can easily grasp the position occupied by the specific company or the like in the field of the prediction target technology by using this marker 110.

For example, the marker 110A is located in the most past map figure M-2. Therefore, it can be estimated that the company of the marker 110A is behind in the technical development in the technical field indicated by the map 100 or has already withdrawn from the market. On the contrary, the marker 110D is located in the most future map figure M + 2. Therefore, it can be determined that the company with the marker 110D is the company with the most advanced technology development. However, it takes some time for the result of technological development to be created as document information, and for this document information to be disclosed and stored in the document DB 31. Accordingly, the map 100 does not strictly indicate the latest state of the technological development status. However, it is possible to infer the presence or absence of development ability of a specific company from the transition of the map figure and the position of the marker 110.
Here, as described above, the map figures M-2, M-1, M0, M + 1, and M + 2 indicate the main areas in the generation year, respectively, and there is literature information and keywords on the outside of each map figure. Exists. For example, even in the generation year of the oldest map graphic M-2, there is a possibility that some document information exists outside the map graphic M-2 (for example, near the marker 110D).
Therefore, when attention is paid to a specific document creator (company, inventor, etc.) in the generation year of the map graphic M-2, the position of the center of gravity of the specific document information group created by the specific document creator (marker 110) ) May be located outside the map figure M-2. When the position indicated by the specific marker 110 exists in the area of the future map figure M + 1, M + 2, the specific document creator can determine that the position is advanced. On the other hand, if the position indicated by the specific marker 110 deviates from any region of the future map graphic M + 1, M + 2, the specific document creator is off the mark in the generation year of the map graphic M-2. It can be judged that the research and development were conducted.

  FIG. 3 is a configuration explanatory diagram showing an outline of the hardware configuration of the document information analysis apparatus 1 and the client terminal 2. As described above, the document information analysis apparatus 1 can be configured as a server computer or the like.

  The document information analyzer 1 includes, for example, a communication interface (in the figure, the interface is abbreviated as “I / F”) 1001, a CPU (Central Processing Unit) 1002, a ROM (Read Only Memory) 1003, and a RAM (Random Access). Memory) 1004 and an auxiliary storage device 1005 can be provided.

  In the auxiliary storage device 1005, for example, in addition to the OS (Operating System), a document database 31, a word database 32, an index database 33, a prediction processing program 1100, a display control program 1110, and a principal component analysis program 1120. The structural analysis program 1130, the search program 1140, and the web server program 1150 can be stored.

  As described above, the document database 31 stores document information such as patent publications in advance. In the word database 32, for example, words, synonyms, synonyms, and the like inappropriate for keywords such as particles and conjunctions are stored in advance. The keywords included in each document information can be extracted by using the document database 31 and the word database 32, respectively. The index database 33 manages the location of keywords for each piece of document information extracted in this way. Therefore, by using the index database 33, necessary document information can be quickly retrieved from the document database 31 storing a large number of document information.

  The prediction processing program 1100 is a program for creating a prediction map figure M + n. The display control program 1110 is a program that performs drawing processing of the map 100 and the like. The principal component analysis program 1120 is a program for performing principal component analysis. The structure analysis program 1130 is a program for analyzing the structure of text data based on a technique such as text mining. The search program 1140 is a program that searches the literature database 31 and the like based on the input search conditions. The web server program 1150 is a program that realizes a web server function.

  The configuration of the client terminal 2 will be described. The client terminal 2 is connected to the document information analysis apparatus 1 via a communication network CN such as the Internet. For example, the client terminal 2 can be configured to include a communication interface 2001, a CPU 2002, a ROM 2003, a RAM 2004, and an auxiliary storage device 2005. is there.

  In addition, the above structure is an illustration and this invention is not limited to this. For example, it may be possible to realize at least a part of each function by using a hardware circuit such as a programmable logic device instead of the program.

  A document information analysis method using the document information analysis apparatus 1 (hereinafter also referred to as the analysis apparatus 1) will be described. FIG. 4 is a flowchart showing an overview of map control processing for generating and outputting the map 100. The flowcharts described below are also the same, but each flowchart shows an outline of the process and is different from an actual program.

  First, the user inputs search conditions via the user interface of the client terminal 2 (S10). The search condition can be performed, for example, by specifying a keyword such as “photocatalyst” or by specifying a patent classification code, a publication date of a document, or the like.

  When acquiring the search condition from the client terminal 2 (S11), the analysis apparatus 1 searches the index database 33 and the document database 31 based on the search condition, and extracts all document information that matches the search condition (S12). More specifically, the index database 33 is used to extract document information having a keyword designated by the user. When document information having the specified keyword is specified, the document information is read from the document database 31. The extracted result is transmitted from the analyzer 1 to the client terminal 2 (S13) and displayed on the screen of the client terminal 2 (S14).

  The user confirms the total number of extracted documents, document names, and the like via the screen of the client terminal 2, and gives approval to the extraction result (S15). If the user is not satisfied with the extraction result, the search condition can be changed and the search can be instructed again.

  When the analysis device 1 confirms the user's approval, the analysis device 1 extracts a keyword from each extracted document information (S16). The keyword extraction result is transmitted from the analysis apparatus 1 to the client terminal 2 (S17) and displayed on the screen of the client terminal 2 (S18). The user confirms whether the desired keyword is included in the keyword extraction result, and gives approval (S19). If the user is not satisfied with the keyword extraction result, it is possible to request another keyword extraction.

  Upon confirming the approval of the user, the analysis device 1 calculates the coordinates of each extracted document information by performing principal component analysis (S20). Subsequently, the analyzer 1 calculates the distribution density of the document information based on the coordinates of each document information (S21).

  The analysis apparatus 1 calculates the coordinates of each extracted keyword by performing principal component analysis (S22). And the analyzer 1 produces | generates the map 100 and transmits to the client terminal 2 (S23). The user confirms the map 100 displayed on the screen of the client terminal 2 (S24).

  FIG. 5 is an explanatory diagram schematically showing how the map 100 is generated. The analysis apparatus 1 calculates the coordinates of each document information and stores them in the document coordinate management table T1. In the document coordinate management table T1, for example, the document number for specifying document information is associated with the X-axis coordinate and the Y-axis coordinate of the document information. Further, the analysis apparatus 1 calculates the coordinates of each extracted keyword (main keyword) and stores it in the keyword coordinate management table T2. For example, the keyword coordinate management table T2 stores a keyword, a document number including the keyword, and the X-axis coordinate and the Y-axis coordinate of the keyword. In the figure, only one document number including a keyword is shown, but a list of all document numbers including the keyword can be included.

  As shown in the lower side of FIG. 5, for example, a large number of block regions 130 can be set on the map 100 by dividing each of the X-axis method and the Y-axis direction of the map 100 by a plurality. The analyzer 1 calculates the distribution density of the document information by calculating the number of document information located in each block region 130. The contour line 101 is set according to the calculated distribution density.

  FIG. 6 is a flowchart showing the prediction process. First, the user inputs a prediction condition to the analysis apparatus 1 via the client terminal 2 (S31). This prediction condition includes, for example, a prediction target period, designation of a prediction model to be used, and the like.

  First, a user interface for specifying a prediction condition will be described with reference to FIG. The prediction condition setting screen P1 shown in FIG. 7 includes, for example, a search formula P11, a target period P12, a target document type P13, a prediction period P14, a prediction unit P15, a prediction model P16, and an output form P17. Can be provided.

  The search expression input unit P11 is for the user to specify search conditions for narrowing down the technology to be predicted. The target period input unit P12 specifies the time when the document information group that is the basis of the map 100 is created.

  The target document type input unit P13 is for specifying which type of document information is the target among the respective document information such as published patent gazettes, scientific and technical papers, technical journals, published technical bulletins, etc. Is. Multiple types of documents can be selected.

  The prediction period input part P14 designates how far the future should be predicted. The lower limit of the forecast period is the current date, and there is no need to set the upper limit. However, the upper limit value can also be set from the relationship between the document information storage status and the prediction accuracy.

  It is also possible to link the selection of the prediction model and the upper limit value of the prediction period. For example, for each prediction model, an upper limit value of a predictable time is set in advance, and the upper limit value may be displayed on the screen P1 according to the selected prediction model. Or the selection of a prediction model, the accumulation | storage state of literature information, and the upper limit of a prediction period can also be linked. For example, if the upper limit value of the prediction period is set as an initial value for each prediction model, and the initial upper limit value is adjusted based on the number of document information groups that match the search formula P11 and the generation time, etc. Good. Such prediction period setting control can be realized by the prediction unit 20 so as to be easily understood by those skilled in the art.

  The prediction unit input unit P15 designates the temporal resolution of prediction. The user may freely specify an arbitrary period, or the user may select from units prepared in advance as shown in FIG. It is also possible to link the specification of the prediction period and the specification of the prediction unit. For example, a predetermined ratio of the prediction period may be selected as a prediction unit. That is, when the prediction period is 10 years from 2005 to 2015, the prediction unit can be displayed as a prediction unit such as 1 year (10%) or 2 years (20%).

  The prediction model input part P15 designates a model used for prediction. The user designates at least one of a plurality of prediction models prepared in advance. If the user does not specify, the prediction model set as the initial value is used.

  The output form input part P17 designates the output form of the prediction result. Examples of output forms include individual display, animation display, thumbnail display, and the like for each prediction unit. The user can appropriately switch the output form even after the prediction process is completed. For example, even when animation display is designated as the first output form, individual map graphics for each prediction unit can be switched and displayed, or all map graphics can be displayed as a list with thumbnails or the like.

  Returning to FIG. When each value input via the prediction condition setting screen P1 is acquired (S32), the analysis apparatus 1 sets a prediction model used for the prediction process (S33). The analysis apparatus 1 generates the reference map graphic M0 and each past map graphic Mn based on the document information group designated by the target period input unit P12 (S34).

  The analysis apparatus 1 generates a predicted map graphic M + n based on the locus of the center of gravity of each map graphic M0, M-n and the prediction model set in S33 (S35). The analysis apparatus 1 provides the client terminal 2 with the map 100 including the map figures Mn, M0, and M + n (S36).

  When the client terminal 2 displays the map 100 (S37), the user can grasp past and future trends of the prediction target technology from the map 100. The user can also designate a specific document creator via the client terminal 2 (S38). This particular document creator can be referred to as the target creator.

  The analysis apparatus 1 extracts the document information created by the target creator from the document information included in the map 100 output in S36, and the center of gravity position (in FIG. 8) is extracted from the coordinates of the extracted document information. GT) is calculated (S39). The literature information group concerning this target creator can also be called a target literature information group.

  The analysis apparatus 1 determines the advanced level of the target creator based on the relative positional relationship between the gravity center position of the target document information group and each map figure M−n, M0, M + n (S40), and the determination result Is output (S41). The user can infer the position of the target creator and the trend of research and development in the technical field to be predicted based on the advancedness determination result.

  The prediction method and the advancedness determination will be described with reference to FIG. The analysis apparatus 1 calculates the center-of-gravity positions G0, G-2, and G-1 of the reference map figure M0 and the past map figures M-2 and M-1, and obtains the locus of each center-of-gravity position. Then, the gravity center positions G + 1 and G + 2 of the future map figures M + 1 and M + 2 are calculated on the assumption that they exist on the extension of the gravity center locus from the present to the past. Such a prediction method will be effective in a technical field formed by stacking a large number of elemental technologies such as the electric field.

  And the line which connects these each gravity center position G0, G-2, G-1 can be displayed as a course display which shows the direction of technical development. Each map figure is associated with a time display 103.

  A map graphic MT based on the target document information group of the target creator can also be displayed on the map 100. This map graphic MT can also be called a target map graphic. Instead of displaying the target map graphic MT, only the center of gravity position GT may be displayed. The gravity center position GT of the target map graphic MT corresponds to the marker 110 shown in FIG.

  The advancedness determination can be performed by various methods. For example, as one method, as shown in FIGS. 2 and 8, a line connecting the center of gravity position GT of the target map figure MT and the center of gravity positions G-2, G-1, G0, G + 1, G + 2 and each map figure M By visualizing the positional relationship between −2, M−1, M0, M + 1, and M + 2, the determination result can also be shown. Further, for example, the relative positional relationship can be digitized and displayed so that the value increases toward the head of the route display.

  As another method, a predetermined score is assigned to each map figure M-2, M-1, M0, M + 1, M + 2 in advance, and each document is determined according to the coordinates of each document information constituting the target document information group. The score can be totaled for each information, and the average value of the total score can be used as a result of the advanced determination.

  Consider a case where SP1 points are assigned to the past map figure M-n, SP2 points are assigned to the current map figure M0, and SP3 points are assigned to the future map figure M + n (SP1 <SP2 <SP3). As described above, since each map figure M-n, M0, M + n is continuously arranged with the adjacent map figures overlapping, the document information constituting the target document information group is one or more maps. May be contained within a shape.

  Here, when the coordinates of the document information constituting the target document information group exist in the area of each map figure Mn, M0, M + n, the score SPt of the document information is the sum of SP1, SP2, and SP3. (SPt = SP1 + SP2 + SP3). In this way, by collecting the scores for each document information constituting the target document information group and obtaining the average value of the scores, it can be used as an advanced index (determination result). In the case of the above example, the larger the value, the more advanced it can be determined.

  Since the present embodiment employs the above-described configuration, the following effects can be obtained. In the present embodiment, a future map figure M + n is created based on the reference map figure M0 and past map figure M-n and the prediction model, and guidance information (103, 104) indicating the course and time of the map figure is associated. The map 100 is created. Therefore, the user can grasp when and how the prediction target technology changes, that is, when and in which area, and when and withdraw from which area, and the convenience of the map 100 is improved.

  In this embodiment, the configuration is such that a prediction model used for prediction processing can be selected from a plurality of prediction models prepared in advance. Therefore, for example, an appropriate prediction model can be selected according to the nature of the prediction target technology, the length of a period for which prediction is desired, the expected prediction accuracy, and the like, and usability is improved.

  In the present embodiment, a plurality of map figures from the past to the future or from the present to the future are continuously displayed in a superimposed manner. Therefore, it is possible to easily confirm a change in technical trend without displaying a special arrow line or the like.

  In this embodiment, the prediction target period is divided into predetermined periods, and map figures M + n and M−n corresponding to the period are generated based on literature information created within each predetermined period. Therefore, since the difference in the center of gravity of each map figure can be made relatively large, it is easy to grasp the transition of technology trends and the usability is improved.

  In the present embodiment, the target map graphic MT based on the target document information group and each map graphic MN, M0, M + N are compared, and the comparison result is output. Therefore, in the prediction target technology, it is possible to grasp how the target creator focused on by the user is positioned, and the usability is improved.

  Based on FIG. 9, a second embodiment of the present invention will be described. Each embodiment described below corresponds to a modification of the first embodiment. In this embodiment, the map 100 is generated by giving a predetermined weight to the document information.

  FIG. 9 is a flowchart showing map control processing according to the present embodiment. This flowchart shares many steps with the flowchart shown in FIG. Therefore, the description that overlaps is omitted, and the characteristic step S21A will be mainly described. In the present embodiment, the predetermined reference information is given a predetermined weight in the reference information group extracted in step S12. The distribution density of is calculated.

  Here, the predetermined document information is, for example, highly utilized document information, and may include document information with a high number of citations, document information with a large number of browsing requests, and a large number of information provisions. Whether the document is related to a pioneer technique or a document related to an improved technique can be determined based on the number of times cited. Moreover, it is considered that a document with a large number of requests for browsing and a large number of information provisions is a document with a high degree of attention around.

  Therefore, since such highly utilized document information is considered to be an important document that occupies an important position in the technical field, the map 100 is created by weighting. Thereby, it is possible to obtain the map 100 reflecting the substantial contents of the prediction target technology, not just the number of documents. Thus, by arranging the predicted map figure M + n based on the map 100 reflecting the actual technical value, it is possible to grasp the technical trend more accurately.

  A third embodiment of the present invention will be described with reference to FIGS. In this embodiment, when displaying the document information on the map 100, the search range is expanded until the preset number of documents is reached.

  By specifying a desired position on the map 100, the user can display all or part of the specified position and the literature information existing in the vicinity thereof on the terminal screen. When the user selects an arbitrary position on the map 100 and requests display of document information (S51), the analysis apparatus 1 extracts all document information within a predetermined range from the designated position (S52). For example, all the document information existing within a predetermined distance from the designated position is selected.

  The analysis apparatus 1 determines whether or not the number of the extracted document information has reached a predetermined number N (S53). If the number has reached (S53: YES), the analysis apparatus 1 sends the information about the extracted document information to the client terminal 2. (S55). For example, a list of document names of extracted document information may be presented, or a list of bibliographic items (creator, creation date, summary, etc.) of the extracted document information may be presented. Or the structure which presents the whole content of the extracted literature information may be sufficient. The client terminal 2 displays the information transmitted from the analysis device 1 on the terminal screen (S56).

  On the other hand, when the number of document information extracted in S52 does not reach the predetermined number N (S53: NO), the analysis apparatus 1 expands the search range and extracts document information again (S54). When a predetermined number N of document information is extracted (S53: YES), the analysis apparatus 1 presents information regarding the extracted document information to the client terminal 2 (S55).

  Thus, by controlling the coordinate range for acquiring the document information, as shown in FIG. 11, a predetermined number N of document information lists P2 can always be displayed, and the usability is improved. Note that it is preferable to generate the document information list P2 such that document information having coordinates closer to the position designated by the user is displayed higher.

  A fourth embodiment of the present invention will be described with reference to FIGS. In this embodiment, a difference between map figures is extracted and displayed on the map 100. As shown in the configuration diagram of FIG. 12, the map control unit 10 is provided with a difference detection unit 18.

  FIG. 13 is a flowchart showing the difference detection process. When the user designates a plurality of map figures to be compared (S61), the analysis apparatus 1 acquires information on each of the designated map figures (particularly information on the center of gravity position and the contour) (S62). Then, the analysis apparatus 1 calculates the difference between the map figures (S63) and outputs the calculated difference to be displayed on the map 100 as a figure (S64). The client terminal 2 displays the difference between the map graphics on the map 100 (S65).

  FIG. 14 is an explanatory diagram schematically showing the difference detection process. For example, the user requests the visualization of the difference by specifying the target map graphic MT and the predicted map graphic M-2. Thereby, as shown in the difference display screen P3, the differences Δ1 and Δ2 between the two map figures are displayed on the map 100, respectively. The difference Δ1 indicates a technical area included only in the predicted map graphic M-2, and the difference Δ2 indicates a technical area included only in the target map graphic MT. The white area where both map figures MT and M-2 match indicates a technical area common to both.

  Therefore, the user can confirm at a glance the matching area and the different area between the map graphics, and the usability is improved. Not only the comparison between the target map figure MT and the predicted map figure M + n, but also a comparison between the reference map figure M0 and the predicted map figure M + n, a comparison between the past map figure M-n and the predicted map figure M + n, etc. can be performed.

  FIG. 15 corresponds to a further modification of the fourth embodiment. In this modification, an average map 100 of a specific industry is generated so that a difference with a specific company belonging to the industry can be detected.

  For example, by using corporate group classification information that classifies a corporate group for each industry, such as the Japanese standard industry classification or the Nomura 400 classification, it is possible to specify the corporate group belonging to each industry. Therefore, when investigating the technical trend of a certain industry as a whole, the user designates a desired industry and requests the analysis apparatus 1 to perform a search. The analysis apparatus 1 can identify a group of companies belonging to the designated industry, extract all the literature information created by each group of companies, and generate the map 100.

  FIG. 15 is a flowchart showing a difference detection process for comparing an average map graphic of a specific industry and a map graphic of a specific company. When the user specifies a specific company (S71), the analysis apparatus 1 uses the company information classification information to specify the industry to which the specified company belongs and the identification code of each company belonging to that industry (S72). .

  The analysis apparatus 1 generates a map graphic for the industry to which the designated company belongs in the same manner as described in the first embodiment (S73). Since this map figure is based on a literature information group created by each company belonging to the industry, it shows the average technology spread of the industry. Therefore, this map graphic can be called an industry average map graphic.

  Subsequently, the analysis apparatus 1 generates a map graphic relating to the designated company based only on the literature information relating to the company designated in S71 (S74). This map figure can be called a specific company map figure, for example. Then, as described above, the analysis apparatus 1 calculates the difference between the industry average map graphic and the specific company map graphic (S75), and causes the client terminal 2 to output the difference (S76). The client terminal 2 displays this difference graphic on the terminal screen (S77).

  As a result, the user can easily grasp the position of the particular company to be noted in the technology development competition of the industry. Further, by generating a predicted map graphic of the industry average map graphic and extracting the difference between the industry average predicted map and the specific company map, the future potential of the specific company can be considered.

  The present invention is not limited to the above-described embodiment. A person skilled in the art can make various additions and changes within the scope of the present invention. For example, those skilled in the art can appropriately combine the above embodiments.

It is explanatory drawing which shows the whole literature information analyzer which concerns on embodiment of this invention. It is a functional block diagram of a literature information analysis device. It is explanatory drawing which shows the hardware and software structure outline | summary of a literature information analyzer. It is a flowchart which shows the process which produces | generates and displays a map. It is explanatory drawing which shows the relationship between a literature coordinate management table and a keyword coordinate management table, and a map. It is a flowchart which shows a prediction process. It is explanatory drawing which shows a prediction condition setting screen. It is explanatory drawing which shows a prediction method etc. typically. It is a flowchart which shows the map control process performed by the literature information analyzer which concerns on 2nd Example. It is a flowchart which shows the literature display control process performed by the literature information analyzer which concerns on 3rd Example. It is explanatory drawing which shows a mode that a predetermined number of literature information list is shown. It is a functional block diagram of the literature information analyzer which concerns on 4th Example. It is a flowchart which shows a difference detection process. It is explanatory drawing which shows a difference detection method typically. It is a flowchart which shows the difference detection process which concerns on the modification of 4th Example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Document information analysis apparatus, 2 ... Client terminal, 10 ... Map control part, 11 ... Map generation condition input reception part, 12 ... Document extraction part, 13 ... Document coordinate calculation part, 14 ... Keyword extraction part, 15 ... Keyword coordinate Calculation unit, 16 ... map generation unit, 17 ... map display unit, 18 ... difference detection unit, 20 ... prediction unit, 21 ... prediction condition input reception unit, 22 ... prediction model setting unit, 23 ... prediction processing unit, 24 ... basis Data storage unit, 30 ... information storage unit, 31 ... document database, 32 ... word database, 33 ... index database, 100 ... map, 101A, 101B, 101C ... contour, 102 ... keyword, 103 ... time display, 104 ... course display 110A, 110B, 110C, 110D ... marker, G, GT ... center of gravity of map figure M ... map figure, M0 ... standard -Up figure, M-n ... past map graphics, M + n ... prediction map graphics, MT ... target map figure, P1 ... prediction condition setting screen

Claims (12)

A map control unit that generates a predetermined map based on map generation conditions;
A prediction unit for predicting a time change of the map generated by the map control unit,
The map control unit
A document information storage unit for storing a plurality of document information each of which is an electronic patent document or academic paper,
By performing principal component analysis based on a combination of a plurality of predetermined keywords extracted from a predetermined group of document information determined by the map generation condition and the number of appearances among the stored document information, the predetermined document A document coordinate calculation unit for calculating the coordinates of each document information of the information group, and
A keyword coordinate calculation unit that calculates the coordinates of each of the predetermined keywords by performing principal component analysis based on the total number of document information including the predetermined keywords and the number of appearances of the predetermined keywords;
A distribution density of each document information is calculated based on the coordinates of each document information calculated by the document coordinate calculation unit, and a map figure having an outline based on the calculated distribution density and the predetermined keyword are visualized respectively. A map generation unit for generating map information,
With
The prediction unit
Based on the first map graphic generated by the map control unit based on the predetermined document information group and the set prediction model, the second map graphic is predicted at a predetermined time point in the future than the first map graphic. A prediction processing unit to
With
The map information is displayed by combining the first map graphic and the second map graphic, and guidance information for guiding the change from the first map graphic to the second map graphic. Is a document information analysis apparatus including at least information on a course from the first map graphic to the second map graphic.   The literature information analysis device according to claim 1, wherein the guidance information further includes information indicating a time when the first map graphic changes to the second map graphic.   The document according to claim 1, wherein the prediction unit includes a prediction model setting unit that selects and sets at least one prediction model specified by a prediction condition from a plurality of prediction models prepared in advance. Information analysis device.   The prediction unit predicts the second map graphic for each of a plurality of future predetermined points, and each of the second map graphics is displayed on the coordinate space where the first map graphic is displayed together with the first map graphic. The document information analysis apparatus according to claim 1, wherein each is visualized and arranged.   The prediction unit divides the predetermined document information group every predetermined period, and the map control unit generates the first map graphic based on the document information group for each predetermined period, respectively. The literature information analysis apparatus according to claim 1, wherein the map information is generated by visualizing a graphic together with the second map graphic.   The prediction unit determines the position of the second map graphic based on the locus of the center of gravity of the first map graphic for each predetermined period, the shape change of the first map graphic, and the set prediction model. The literature information analysis apparatus according to claim 1, wherein at least one or more shapes are predicted.   The literature information analysis apparatus according to claim 1, wherein both a linear prediction model and a nonlinear prediction model are prepared in advance as the prediction model, and at least one of the prediction models can be set.   The prediction unit uses the map control unit to generate a predetermined map figure based on a document information group created by a selected predetermined document creator among the document information groups stored in the document information storage unit. The literature information analysis apparatus according to claim 1, wherein the reference map graphic is generated, the predetermined map graphic is compared with the first map graphic and the second map graphic, and the comparison result is output.   The map control unit performs weighting so that a document information group having a higher degree of use in the predetermined document information group has a higher score, and calculates the coordinates of each document information in the predetermined document information group, respectively. The literature information analysis apparatus according to claim 1.   The literature information analyzer according to claim 1 provided with the difference detection part which detects the difference between a plurality of map figures created by said map control part. Computer
Map control means for generating a predetermined map based on specified map generation conditions;
A program that functions as a prediction unit that predicts a time change of a map generated by the map control unit,
The map control means includes
Predetermined to be extracted from a predetermined group of document information determined by the map generation condition out of each document information stored in a document information storage unit that stores a plurality of document information each of which is an electronic patent document or academic paper. Document coordinate calculation means for calculating the coordinates of each document information of the predetermined document information group by performing principal component analysis based on the combination and the number of appearances of a plurality of keywords,
Keyword coordinate calculation means for calculating the coordinates of each of the predetermined keywords by performing principal component analysis based on the total number of document information including the predetermined keywords and the number of appearances of the predetermined keywords,
A distribution density of each document information is calculated based on the coordinates of each document information calculated by the document coordinate calculation means, and a map figure having an outline based on the calculated distribution density and the predetermined keyword are visualized respectively. Map generation means for generating map information,
Are provided as internal functions,
The prediction means includes
A prediction model setting means for selecting and setting any one prediction model from a plurality of prediction models prepared in advance based on the specified prediction condition;
Based on the first map graphic generated by the map control means based on the predetermined document information group and the set prediction model, a second map graphic at a predetermined time point in the future than the first map graphic is obtained. Prediction processing means for predicting;
Are provided as internal functions,
The map information includes the first map graphic and the second map graphic, and information indicating a course and a time from the first map graphic to the second map graphic. program. Receiving a prediction model, a prediction target period, and a prediction condition including at least one keyword;
Extracting a predetermined document information group including a keyword in the prediction condition from a document information group that is a patent document or an academic paper held so far;
Generating a first map graphic for each predetermined period based on the predetermined document information group belonging to each predetermined period for all the periods from the predetermined predetermined time point to the present time indicated in the prediction target period; ,
Generating a second map graphic at a predetermined time in the future based on the prediction model specified by the prediction condition and each generated first map graphic; and
Combining each of the first map graphic and the second map graphic, and guide information for guiding a change from the first map graphic to the second map graphic, and displaying them on the same screen, respectively;
Including
The step of generating the first map graphic includes:
Calculating the coordinates of each document information of the predetermined document information group by performing principal component analysis based on a combination of a plurality of predetermined keywords and the number of appearances included in the extracted predetermined document information group And calculating the distribution density of each document information based on the calculated coordinates of each document information, and generating a map figure based on the calculated document density,
The step of generating the second map graphic includes:
The second map figure is generated by calculating the centroid position of each first map figure and inputting the locus of each centroid position into the prediction model.
Literature information analysis method.

Images