JP6488399B2 - Information presentation system and information presentation method - Google Patents

Description

  The present invention relates to an information presentation system and an information presentation method.

  Smartphones, tablet terminals, and information terminals such as interactive robots have become familiar, and systems that provide useful information and services to users are increasing due to the interaction between users and systems.

  For example, the information presentation system holds in advance provision information that an administrator or the like wants to provide to the user. When the information presentation system receives an input of a search query from the user, the information presentation system presents the provided information, which is estimated to be of interest to the user, together with search results such as a web page and an image, as useful information to the user. In addition, for example, the information presentation system holds the association between the provider of the provision information and the keyword in advance, and presents the provision information of the provider corresponding to the keyword to the user who has input the keyword as a search query.

  As a background art of this technical field, there is JP-A-2004-102996 (Patent Document 1). This gazette states that “the attribute of transition from the migration source web page to the migration destination web page is stored in the history database 21 as an arc attribute, and the arc attribute of the path to the web page that can be migrated from any web page is stored in the database 21. Read, compare this arc attribute with the user search condition, recommend a Web page candidate to be transferred next, correct the user search condition with the arc attribute to the Web page selected from the candidate, compare processing, recommendation processing, The presentation process and the correction process are repeated, and finally the corrected user search condition is fed back to the arc attribute of all routes, and the database 21 is updated with this arc attribute "(see summary).

JP 2004-102996 A

  The technique described in Patent Literature 1 cannot present to a user provided information that is not associated with a search query or a search result input by the user in advance. Although the search query input by the user indicates the user's interest at the time of input, the user's potential interest may be broader than what is described in the search query.

  Therefore, even provided information that is not directly related to the search query may fall under the potential interest of the user. In this case, the provided information is highly useful information for the user. In addition, it is desirable for the provider to present the provision information not only to users who are directly interested in the provision information, but also to all users who are potentially interested.

  However, if the information presentation system presents provided information that is not related or directly related to the search query together with the search result based on the search query, even if the provided information is a potential concern of the user There arises a problem that the user is unlikely to browse the provided information. This is because the obvious interest of the user at the time of receiving the search result information is the search query and information directly related to the search query, and the potential interest of the user to the provided information is sufficiently aroused. It is thought that there is not.

  In view of this, an object of one embodiment of the present invention is to induce user's potential interest by guiding the user's interest to provided information to be presented, and present useful information for the user.

  In order to solve the above problems, one embodiment of the present invention employs the following configuration. An information presentation system for presenting information to a user, comprising a processor and a storage device, wherein the storage device receives a first search query received from the user, and information collected from a database based on the first search query A first collection information group comprising: a predetermined collection information; a second search query linked to the provision information; and a second collection information group comprising collection information from the database based on the second search query And the processor includes the first search query and the word included in the first collection information group as vertices, and vertices indicating words included in the same collection information are connected by edges. The first graph is generated, and the second search query and the word included in the second collection information group are set as vertices, and the vertices indicating the words included in the same collection information are connected by edges. In addition, a second graph is generated, an integrated graph is generated by integrating the first graph and the second graph at a common vertex of the first graph and the second graph, and information is presented to the user. A first process is executed, and in the first process, a vertex on the first path connecting the first vertex indicating the first search query and the second vertex indicating the second search query on the integrated graph. When the vertex other than the first vertex is selected and the second vertex is selected, the provided information is presented to the user, and when the vertex other than the second vertex is selected, the selected vertex One or more pieces of collected information including a word indicating the above are selected from the first collection information group and the second collection information group, and at least a part of the selected one or more pieces of collection information is presented to the user. Information presentation system.

  According to one embodiment of the present invention, it is possible to arouse a user's potential interest in provided information and present useful information for the user.

It is a block diagram which shows the structural example of the information presentation system in Example 1. FIG. 3 is a flowchart illustrating an operation example of the information presentation system in the first embodiment. 6 is a flowchart illustrating an example of a graph expansion process in the first embodiment. 4 is an example of first graph information in the first embodiment. FIG. 6 is a diagram illustrating an example of a process for integrating a first graph and a second graph in the first embodiment. FIG. 10 is a diagram illustrating an example of vertex selection processing on a path from a first vertex to a second vertex in the first embodiment. It is an example of the vertex selection order on the path | route from the 1st vertex to the 2nd vertex in Example 1. FIG. It is a figure which shows an example of the transition of the presentation information in Example 1. FIG. It is a figure which shows an example of the transition of the presentation information in Example 1. FIG. 6 is a diagram illustrating an example of a first graph expansion process in Embodiment 1. FIG. FIG. 10 is a diagram illustrating an example of a second graph expansion process according to the first embodiment. 3 is an example of an integrated graph in Example 1. FIG. It is a block diagram which shows the structural example of the information presentation system in Example 2. FIG.

  Embodiments of the present invention will be described below with reference to the accompanying drawings. This embodiment is merely an example for realizing the present invention, and does not limit the technical scope of the present invention.

  FIG. 1 shows a configuration example of an information presentation system 100 according to the first embodiment of the present invention. The information presentation system 100 is a system that receives a search query input from a user, collects related information, and presents it to the user. The information presentation system 100 includes, for example, a computer 110, a network 116, an information collection server 117, databases 120 to 122, and an advertisement information database 140.

  The information collection server 117 performs a search from the databases 120 to 122 using the search query received from the computer 110 and transmits the collected search results to the information presentation system 100. The advertisement information database 140 holds advertisement information.

  The computer 110 includes a user interface 111, a processor 112, a storage 113, a memory 114, and a network interface 115. The user interface 111 includes a device that receives input from the user, such as a keyboard and a mouse. The user interface 111 includes a device such as a display device or a printer that outputs the execution result of the program in a format that can be viewed by the user.

  The processor 112 executes a program stored in the program storage unit 130 of the memory 114. The memory 114 includes a ROM that is a nonvolatile storage element and a RAM that is a volatile storage element. The ROM stores an immutable program (for example, BIOS). The RAM is a high-speed and volatile storage element such as a DRAM (Dynamic Random Access Memory), and temporarily stores a program executed by the processor 112 and data used when the program is executed.

  The storage 113 is a large-capacity non-volatile storage device such as a magnetic storage device (HDD) or a flash memory (SSD), and stores programs executed by the processor 112 and data used when the programs are executed. That is, the program is read from the storage 113 or the memory 114, loaded into the memory 114, and executed by the processor 112.

  The storage 113 stores input collection information 141 and advertisement collection information 142. Note that some or all of the programs and data stored in the memory 114 may be stored in the storage 113, or some or all of the programs and data stored in the storage 113 are stored in the memory 114. It may be.

  The network interface 115 is a network interface device that controls communication with other devices in accordance with a predetermined protocol. The network interface 115 includes a serial interface such as USB.

  The program executed by the processor 112 is provided to the computer 110 via a removable medium (a computer-readable non-transitory storage medium such as a CD-ROM or a flash memory) or the network 116, and the non-temporary storage medium May be stored in the non-volatile storage 113. For this reason, the computer 110 may have an interface for reading data from a removable medium.

  The information presentation system 100 is a computer system configured on one computer physically or on a plurality of computers configured logically or physically, and operates on separate threads on the same computer. Alternatively, it may operate on a virtual machine constructed on a plurality of physical computer resources.

  The program storage unit 130 is a graph information generation unit 131, a presentation information generation unit 132, a search query acquisition unit 133, an advertisement information acquisition unit 134, a graph integration unit 135, a route selection unit 136, a vertex selection unit 137, and a program, respectively. The advertisement information output unit 138 is stored.

  The processor 112 operates as a functional unit that realizes a predetermined function by operating according to a program. For example, the processor 112 functions as a graph information generation unit by operating according to the graph information generation unit 131, and functions as a presentation information generation unit by operating according to the presentation information generation unit 132. Furthermore, the processor 112 also operates as a functional unit that implements each of a plurality of processes executed by each program. A computer and a computer system are an apparatus and a system including these functional units.

  When the search query acquisition unit 133 acquires a search query from the user via the user interface 111, the search query acquisition unit 133 transmits the search query from the network interface 115 to the information collection server 117 via the network 116. The information collection server 117 executes a search using a search query from the databases 120 to 122 and transmits the search result to the computer 110. A search result transmitted from the information collection server 117 to the computer 110 is stored as input collection information 141 in the storage 113 or the memory 114.

  On the other hand, the advertisement information acquisition unit 134 acquires advertisement information stored in the advertisement information database 140 via the network interface 115 and the network 116. Advertisement information is an example of provided information. The provided information is predetermined information that the administrator of the information presentation system 100 wants to provide to the user.

  Further, the advertisement information database 140 holds a search query associated with each advertisement information, and the advertisement information acquisition unit 134 acquires a search query associated with the acquired advertisement information from the advertisement information database 140. Hereinafter, a search query associated with each advertisement information is referred to as an advertisement search query. Each of the advertisement information includes, for example, the name of a product to be advertised and a URL of a website including information on the product. In addition, for example, position information may be given to each advertisement information. In addition, for example, attribute information such as age and gender, which is an advertisement presentation target, may be assigned to each advertisement information. Further, for example, advertiser information such as the reliability of the advertiser and the advertisement placement fee may be given to each advertisement information.

  The advertisement information acquisition unit 134 transmits the advertisement search query to the information collection server 117 via the network interface 115 and the network 116. The information collection server 117 searches the databases 120 to 122 using an advertisement search query, and transmits the search result to the computer 110. The search result by the advertisement search query transmitted from the information collection server 117 to the computer 110 is stored in the storage 113 or the memory 114 as the advertisement collection information 142.

  The graph information generation unit 131 generates first graph information for the input collection information 141 and generates second graph information for the advertisement collection information 142. The first graph information and the second graph information will be described later. The graph integration unit 135 generates an integrated graph by integrating the first graph included in the first graph information and the second graph included in the second graph information.

  The route selection unit 136 selects a route connecting the vertex indicating the search query input by the user and the vertex indicating the advertisement search query on the integrated graph. The vertex selection unit 137 selects a vertex in the route selected by the route selection unit 136. The presentation information generation unit 132 generates the presentation information using the input collection information 141 and the advertisement collection information 142 according to the vertex selected by the vertex selection unit 137, and outputs the presentation information via the user interface 111.

  FIG. 2 shows an example of the overall processing of the information presentation system 100. The search query acquisition unit 133 acquires a search query that the user interface 111 has received an input from a user, for example (S202). In addition, the search query in a present Example shall be designated with a character string, a moving image, or an image, for example.

  Subsequently, the search query acquisition unit 133 transmits the acquired search query from the network interface 115 to the information collection server 117 via the network 116. The information collection server 117 searches the databases 120 to 122 using the received search query, and collects related information (website, video, image, map, etc.) of the search query, which is a search result based on the search query. To do. The information collection server 117 returns each collected related information to the computer 110 (S203). Hereinafter, each piece of related information collected by the information collection server 117 is referred to as collection information.

  When a search query is specified by a character string, a website including the character string in the text, a website including the character string in meta information, a moving image, an image, a map, and the like are related information of the search query. It is an example. In addition, when a search query is specified by a moving image or image, a moving image, an image, and a map similar to the moving image or the image, and a website including a moving image or image similar to the moving image or the image are searched. It is an example of the relevant information of a query. The computer 110 stores information including each of the collection information received from the information collection server 117 as input collection information 141 in the storage 113 or the memory 114. Details of the input collection information 141 will be described later.

  Subsequently, the graph information generation unit 131 generates first graph information for the input collection information 141 (S204).

  Hereinafter, a generation example of the first graph information by the graph information generation unit 131 will be described. For example, the graph information generation unit 131 acquires a word in each collection information included in the input collection information 141. For example, the graph information generation unit 131 obtains a word by performing morphological analysis on a sentence included in the collected information with reference to a dictionary or the like stored in advance in the storage 113 or the like. In addition, the word in a present Example shows the character string defined in the said dictionary etc., for example, The concept which does not necessarily correspond with the grammatical word may be sufficient. Further, the search query and the advertisement search query input from the user are regarded as one word.

  Note that the graph information generation unit 131 may acquire a word from the text of the collected information, for example, or may acquire a word from the metadata of the collected information. Further, the graph information generation unit 131 may acquire a moving image, an image, and the like from the collected information. Hereinafter, in this embodiment, for convenience of explanation, “word” is a concept including a moving image, an image, and the like.

  The graph information generation unit 131 counts the number of appearances in the input collection information 141 for each acquired word. Further, the graph information generation unit 131 counts the number of occurrences of the pair of words in the same collection information for all pairs of two words, that is, the number of co-occurrence of the pair.

  The graph information generation unit 131 generates a co-occurrence graph in which each acquired word is a vertex, and two vertices indicating pairs of words having a co-occurrence number of 1 or more are connected by an edge. Note that the length of each edge is determined by, for example, a value obtained by substituting the co-occurrence number of two vertices connected by the edge into a predetermined decreasing function. Specifically, for example, the graph information generation unit 131 determines the length of each edge by 1 / (number of co-occurrence). For example, the graph information generation unit 131 determines the number of appearances of the word indicating each vertex in the input collection information 141 as the size of the vertex.

  The graph information generation unit 131 includes the generated graph as the first graph in the first graph information. In addition, the graph information generation unit 131 includes, for example, input collection information 141, a data table indicating the number of appearances of each word, and a data table indicating the number of co-occurrence of each pair of words in the first graph information.

  FIG. 4 shows an example of the first graph information. The first graph information 400 includes, for example, input collection information 401, a first graph 402, a data table 403, and a data table 404. The input collection information 401 is an example of the input collection information 141. The input collection information 401 includes collection information 420 to 422. Each of the collected information 420 to 422 is a website, a moving image, an image, or the like, and includes at least one of the words v1 to v18.

  For example, the collection information 420 includes words v1, v5, and v6. And the 1st graph 402 is produced | generated from the input collection information 401 of FIG. Here, the word v1 is a search query input by the user. Further, the data table 403 shows the data structure of the vertices constituting the first graph 402, and the data table 404 shows the data structure of the edges constituting the first graph 402.

  The data table 403 includes vertices (words) constituting the first graph 402 and the number of appearances in the input collection information 401. The data table 404 includes two words that are end points of the edges constituting the first graph 402 and the number of co-occurrence thereof. Here, the co-occurrence number is the number of two words that appear in the same collection information of the input collection information 401. For example, in the collection information 420, the word pairs v1 and v3, v1 and v6, and v3 and v6, respectively. Origin 1 is counted.

  Returning to the description of FIG. The advertisement information acquisition unit 134 acquires one or more advertisement information and an advertisement search query associated with the one or more advertisement information from the advertisement information database 140 via the network 116 from the network interface 115 and stores them in the storage 113 or the memory 114. (S205).

  Note that the advertisement information acquisition unit 134 may acquire all advertisement information stored in the advertisement information database 140, or may narrow down advertisement information acquired according to a predetermined condition. A specific example of narrowing down will be described. For example, the advertisement information acquisition unit 134 acquires the geographical current position of the computer 110 based on information such as a base station to which the computer 110 is connected, and compares the current position with the position of each advertisement information. Based on this, narrow down the advertisement information to be acquired. Specifically, for example, the advertisement information acquisition unit 134 acquires advertisement information to which a position within a predetermined distance from the position of the computer 110 is given.

  Further, for example, user attribute (age, gender, etc.) information may be stored in the storage 113 or the memory 114 in advance. At this time, the advertisement information acquisition unit 134 may acquire the advertisement information to which attribute information similar to the user is given by comparing the user attribute information with the attribute information given to each of the advertisement information. Further, for example, the advertisement information acquisition unit 134 refers to the advertiser information of each advertisement information, and receives advertisement information from an advertiser whose reliability is a predetermined value or more and / or whose advertisement insertion fee is a predetermined value or more. You may get it.

  The advertisement information acquisition unit 134 transmits the acquired one or more advertisement search queries from the network interface 115 to the information collection server 117 via the network 116. The information collection server 117 collects related information of each advertisement search query from the databases 120 to 122, and returns each collected related information to the computer 110 as collected information associated with each advertisement search query (S206). The computer 110 stores information including the collected information associated with each advertisement search query received from the information collection server 117 as the advertisement collection information 142 of the advertisement information corresponding to the advertisement search query in the storage 113 or the memory 114.

  Subsequently, the graph information generation unit 131 generates second graph information for each advertisement collection information 142 by, for example, the same method as the first graph information generation processing (S207). Here, since the series of processes consisting of steps S205 to S207 is independent of the series of processes consisting of steps S202 to S204, there is no need for synchronization. Moreover, it is preferable that the graph information generation unit 131 generates second graph information using the latest advertisement information. Therefore, it is preferable that the processes in steps S205 to S207 are performed periodically regardless of whether a search query is input from the user.

  Subsequently, the graph integration unit 135 selects one graph information from the one or more second graph information generated in step S207, and selects the first graph and the second graph included in the selected graph information. It is determined whether or not integration is possible (S208). For example, the graph integration unit 135 randomly selects one piece of graph information from the second graph information. For example, when the first graph and the second graph include a vertex indicating a common word, the graph integration unit 135 determines that two pieces of graph information can be integrated, and integrates the graph at the vertex. A specific example of the graph integration process will be described later.

  Further, for example, if the first graph and the second graph do not include a vertex indicating a common word, the graph integration unit 135 determines that the two graph information cannot be integrated (S208: NO), and the connector 1 The process proceeds to the flowchart of FIG. For example, when the process of step S208 is repeated a predetermined number of times, or when a predetermined time has elapsed since the first determination process of step S208, the process of FIG.

  FIG. 5 shows an example of the integration process of the first graph and the second graph. The first graph 501 and the second graph 502 include common vertices v14, v17, and v18. For example, the graph integration unit 135 generates the integrated graph 503 by integrating the first graph and the second graph at the common vertex.

  The graph integration unit 135 calculates the sum of the sizes of the common vertices of the first graph 501 and the sizes of the common vertices of the second graph 502, for example, Is the size of each common vertex in the integrated graph 503. When the first graph 501 and the 502nd graph include a common edge, the graph integration unit 135 determines the number of co-occurrence corresponding to the common edge of the first graph 501 and the second graph 502. The sum of the co-occurrence number corresponding to the common edge is calculated, and the calculated sum is set as the co-occurrence number of the common edge in the integrated graph 503.

  Returning to the description of FIG. When two pieces of graph information can be integrated (S208: YES), the route selection unit 136 changes from the first vertex indicating the search query acquired in S202 to the second vertex indicating the advertisement search query on the integrated graph. A route search is performed, and one route connecting the two vertices is selected (S209).

  Hereinafter, an example of processing in which the route selection unit 136 selects a route connecting the first vertex and the second vertex will be described. For example, the route selection unit 136 calculates the distances of all the routes connecting the first vertex and the second vertex, and selects the route having the shortest distance. In addition, the route selection unit 136 may select a route whose distance is equal to or less than a predetermined threshold, for example. Note that the route selection unit 136 calculates the distance of the route by, for example, substituting the sum of the co-occurrence numbers of edges included in the route into a predetermined decreasing function.

  In addition, the route selection unit 136 has a shortest distance from a route that excludes a route including an edge having a co-occurrence degree between two vertices smaller than a threshold from all routes connecting the first vertex and the second vertex. May be selected. The co-occurrence degree between two vertices is defined by (number of co-occurrence of the two vertices) / (total number of co-occurrence in the integrated graph). The route selection unit 136 can exclude a route directly connecting words having low relevance by excluding an edge connecting two vertices having a low co-occurrence degree.

  Further, the route selection unit 136 may select a route having the shortest distance from routes passing through a predetermined number or more of vertices whose median centrality is equal to or greater than a threshold value. Further, the route selection unit 136 may select a route having the shortest distance from routes passing through a predetermined number or more of vertices whose degree centrality is equal to or greater than a threshold value. The route selection unit 136 can include a vertex indicating a word having a large amount of information in the route by selecting a route including a vertex having a high centrality such as an intermediary centrality or an order centrality.

  Subsequently, the vertex selection unit 137 selects a vertex on the route selected by the route selection unit 136 (S210). For example, in step S210, the vertex selection unit 137 selects a vertex farther from the first vertex in the route than the vertex selected in the previous step S210. That is, the vertex selection unit 137 selects a vertex closer to the second vertex in the path than the vertex selected in the previous step S210.

  Hereinafter, an example of processing in which the vertex selection unit 137 selects a vertex on the route will be described. For example, the vertex selecting unit 137 selects a vertex closest to the first vertex in the route from unselected vertices on the route.

  For example, the vertex selection unit 137 may select, from unselected vertices on the route, a vertex whose vertex size is equal to or larger than a threshold and closest to the first vertex. Further, the vertex selection unit 137 is, for example, a vertex that is closer to the second vertex in the route than the vertex selected in the previous step S210 from an unselected vertex on the route, and has the highest median centrality or A vertex having the largest degree centrality may be selected.

  FIG. 6A shows an example of vertex selection processing on a path from the first vertex to the second vertex. The word v1 of the integrated graph 601 is a word indicating the search query input by the user, the word v32 is a word indicating the advertisement search query, and the path from v1 to v32 is five words in the order of v1-v6-v18-v22-v32. It consists of

  FIG. 6B is an example of a table indicating the order in which the vertex selection unit 137 selects vertices. The vertex selection unit 137 selects vertices according to the selection order indicated by the table 602 when the method of selecting the vertex closest to the vertex indicating the search query is used.

  Returning to the description of FIG. Subsequently, the presentation information generation unit 132 generates presentation information from the input collection information 141 based on the vertices selected by the vertex selection unit 137 and outputs the generated presentation information to the user interface 111 to present to the user. (S211).

  An example of generating the presentation information will be described. The presentation information generation unit 132 acquires collection information from the input collection information 141. Further, the presentation information generation unit 132 acquires advertisement related information of the word from the input collection information 141 or the advertisement collection information 142 based on the word indicated by the vertex selected by the vertex selection unit 137.

  Note that the word indicated by the vertex selected by the vertex selection unit 137 is an example of advertisement-related information of the word. Moreover, when the collection information including the word indicated by the vertex is a website, the title, URL, snippet, and the like of the website are examples of advertisement-related information of the word. When there is a plurality of collected information including a word indicating the vertex selected by the vertex selecting unit 137, the presentation information generating unit 132, for example, has the number of occurrences of the word equal to or greater than a threshold, or the number of appearances of the word is the maximum. The collected information is selected, and advertisement related information is acquired from the selected collected information. The presentation information generation unit 132 includes the acquired collection information and the acquired related information in the presentation information.

  Subsequently, the advertisement information output unit 138 determines whether or not the user has selected the advertisement-related information included in the presentation information, that is, the advertisement-related information selected in step S210, via the user interface 111 (S212). . If the user has not selected the advertisement related information (S212: NO), the process returns to step S202. The selection of the advertisement related information by the user indicates that the user is interested in the advertisement related information. The selection of the advertisement related information by the user indicates that the user has not shown interest in the advertisement related information.

  When the user selects advertisement-related information (S212: YES), the advertisement information output unit 138 determines whether or not the selected advertisement-related information is advertisement information (S213). That is, the advertisement information output unit 138 determines whether or not the vertex selected in step S210 is the second vertex.

  If the advertisement information output unit 138 determines that the advertisement-related information is advertisement information (S213: YES), the advertisement information output unit 138 acquires the advertisement information from the advertisement collection information 142 and outputs it to the user (S214), and ends the process.

  If the advertisement-related information is not advertisement information (S213: NO), there is advertisement-related information that has not been presented to the user, so the advertisement information acquisition unit 134 uses the word indicated by the vertex corresponding to the advertisement-related information as a search query. The information is transmitted from the network interface 115 to the information collection server 117 via the network 116. The information collection server 117 collects related information of the received search query from the databases 120 to 122 and returns the related information to the computer 110 (S215).

  The computer 110 stores information including each received collection information in the storage 113 or the memory 114 as new input collection information 141. And it returns to step S210 and the vertex selection part 137 selects the vertex on a path | route in order to determine the next advertisement relevant information shown to a user.

  In the example of FIG. 6, in the first step S210, the vertex selecting unit 137 selects, for example, a vertex indicating the word v6. When the user selects the advertisement related information of the word v6 in step S212, the vertex selection unit 137 selects v18 which is the next vertex on the route in the next step S210.

  FIG. 7 shows a first example of the transition of the presentation information. The presentation information group 700 is an example of transition of presentation information when the user always selects advertisement-related information in step S212. Further, in the presentation information group 700, the vertices on the route shown in the integrated graph 601 in the example of FIG. 6 are selected in the selection order shown in the table 602, the search query input by the user is v1, and the advertisement search query is v32. This is an example.

  The presentation information 710 is the presentation information generated in the first step S211. The presentation information 710 includes display sections 711 to 713. The display section 711 displays v1 that is a search query input by the user. The display section 712 displays input collection information including the collection information of the search query v1. The display section 713 displays v6 advertisement related information. Here, each of the advertisement related information included in the display section 713 includes, for example, the word v6, the word adjacent to the word v6 in the integrated graph, or the URL of the website collected by the information collection server 117 using the word v6 as a search query and Snippets and the like.

  The presentation information 730 is an example of the presentation information generated in the next step S211 when the user selects any of the advertisement related information included in the display section 713 of the presentation information 710 in the first step S212. The presentation information 730 includes display sections 731 to 733. The display section 731 displays v6 or v6 advertisement related information selected by the user. The display section 732 displays input collection information including collection information with v6 as a search query. The display section 733 displays v18 advertisement related information.

  Note that the display section 732 may display collected information using v1 and v6 as search queries. In this case, the content displayed in the display section 732 is a narrowed search result by the search query v6 for the input collection information displayed in the display section 712.

  The presentation information 750 is an example of presentation information that is finally obtained when the user repeatedly selects advertisement-related information. Presentation information 750 includes display sections 751-752. The display section 751 displays v32 or v6 v32 advertisement related information selected by the user. The display section 752 displays advertisement information. By repeatedly selecting the advertisement-related information, the user can gradually receive the advertisement-related information that is strongly related to the advertisement information.

  FIG. 8 shows a second example of the transition of the presentation information. The presentation information group 800 differs from the presentation information group 700 in that the presentation information 710 includes a display section 812 instead of the display sections 712 to 713, and the presentation information 730 includes a display section 832 instead of the display sections 732 to 733. . The display section 812 and the display section 832 display advertisement-related information mixed in the input collection information. When the presentation information generation unit 132 generates the presentation information in which the advertisement related information is mixed in the input collection information, the user is more likely to select the advertisement related information.

  FIG. 3 shows an example of the graph expansion process executed when the graph integration unit 135 determines in step S208 that the first graph and the second graph cannot be integrated. The graph integration unit 135 extends the first graph (S302). Specifically, the graph integration unit 135 selects an extension vertex from vertices constituting the first graph and other than the first vertex.

  The graph integration unit 135 transmits a word indicating the selected extension vertex as a search query from the network interface 115 to the information collection server 117 via the network 116. The computer 110 receives related information, which is a search result based on a search query indicating an extension vertex, collected by the information collection server 117 and adds it to the input collection information 141. The graph integration unit 135 extends the graph information by adding vertices and edges to the first graph information using the words included in the received related information and the number of co-occurrence in the input collection information 141.

  Hereinafter, an example of a method in which the graph integration unit 135 selects an extension vertex will be described. For example, the graph integration unit 135 may select a vertex having the maximum degree centrality from unselected vertices other than the vertex indicating the search query input by the user. Further, the graph integration unit 135 may select, for example, a vertex having the maximum median centrality from unselected vertices other than the vertex indicating the search query input by the user. Further, for example, the graph integration unit 135 may select a vertex having the longest distance from the vertex indicating the search query input by the user from unselected vertices other than the vertex indicating the search query input by the user. .

  Subsequently, the graph integration unit 135 expands the second graph (S303). Specifically, the graph integration unit 135 selects an extension vertex from vertices constituting the second graph and other than the second vertex. The graph integration unit 135 transmits a word indicating the selected extension vertex as a search query from the network interface 115 to the information collection server 117 via the network 116.

  The computer 110 receives related information, which is a search result by the search query indicating the vertex for expansion, collected by the information collection server 117, and adds it to the advertisement collection information 142. The graph integration unit 135 expands the graph information by adding vertices and edges to the second graph information using the words included in the received related information and the number of co-occurrence in the advertisement collection information 142. After the process of step S303 is executed, the process returns to the process of step 208 in FIG. Note that the graph integration unit 135 may execute only one of step S302 and step S303.

  FIG. 9 shows an example of the expansion process of the first graph. FIG. 9 shows an example in which the graph integration unit 135 selects the word v7 having the maximum degree centrality as the expansion vertex in the first graph 901, and generates the first expansion graph 902 from the related information of the search query v7. It is.

  FIG. 10 shows an example of the expansion process of the second graph. FIG. 10 shows that the graph integration unit 135 is the word having the maximum degree centrality in the second graph 1001 as in the first graph expansion process. In this example, v22 is selected as the vertex for expansion, and the second expansion graph 1002 is generated based on the related information of the search query v22.

  FIG. 11 shows an example of an integrated graph. The integrated graph 1101 is a graph in which the first extended graph 902 shown in FIG. 9 and the second extended graph 1002 shown in FIG. 10 are integrated. The first extended graph 902 and the second extended graph 1002 include a common vertex v42, and the graph integration unit 135 combines the first extended graph 902 and the second extended graph 1002 by combining the vertex v42. A graph 1101 is generated.

  Note that the information presentation system 100 may generate the second graph information using, for example, a search query input by another user with respect to another information presentation system instead of the advertisement search query. By using each user's search query as the second graph information in each other's information presentation system, it is possible to determine whether each of the two users is interested in the other party, that is, matching between people is possible Become.

  As described above, the information presentation system 100 according to the present embodiment selects the vertex on the path connecting the first vertex and the second vertex on the integrated graph in the direction from the first vertex to the second vertex, Advertisement related information corresponding to the selected vertex is presented to the user. Thereby, the information presentation system 100 can guide | invade a user's interest gradually to advertisement information from the search query which the user input. As a result, the information presentation system 100 can raise the user's potential interest in the advertisement information even when the advertisement information is not directly related to the search query entered by the user, and is useful for the user. Information can be presented.

  In this embodiment, an example of an information presentation system including a robot that performs a dialog with a user will be described. A description of the same configuration as that of the information presentation system 100 of the first embodiment is omitted. FIG. 12 shows a configuration example of the information presentation system 1200 of this embodiment. The information presentation system 1200 includes a robot 1210 that interacts with the user and makes suggestions to the user through the interaction. The information presentation system 1200 further includes a network 116, an information collection server 117, a conversation database 1220-1222, and a proposal information database 1240.

  The robot 1210 includes a microphone 1211, a processor 112, a storage 113, a memory 114, a network interface 115, and a speaker 1218. The microphone 1211 is a device that receives voice input from the user. The speaker 1218 is a device that outputs a processing result or the like by the robot 1210 by voice. The storage 113 holds input collection information 1241 and proposal collection information 1242. The input collection information 1241 and the proposal collection information 1242 will be described later.

  The memory 114 includes a program storage unit 130. The program storage unit 130 includes a speech recognition unit 1231, a graph information generation unit 131, an utterance content generation unit 1233, a speech synthesis unit 1239, a proposal information acquisition unit 1234, a graph integration unit 135, a route selection unit 136, and a vertex. The selection unit 137 and the advertisement information output unit 138 are stored. Each of the conversation databases 1220 to 1222 holds a correspondence between the search query and one or more response sentences corresponding to the search query.

  When the microphone 1211 acquires the utterance content by voice input from the user, the acquired utterance content is transferred to the voice recognition unit 1231. The voice recognition unit 1231 acquires text information of the utterance content by executing voice recognition processing on the acquired utterance content. Further, the speech recognition unit 1231 may execute processing such as morphological analysis and semantic interpretation on the text information acquired in the speech recognition processing.

  The voice synthesizer 1239 outputs a response sentence to the utterance content to the user via the speaker 1218. The information presentation system 1200 interactively responds to the voice input from the user by the above-described processing.

  Hereinafter, an example of processing in which the information presentation system 1200 makes a proposal to the user will be described. The voice recognition unit 1231 transmits an utterance search query including at least a part of the text of the user's utterance content from the network interface 115 to the information collection server 117 via the network 116. The information collection server 117 acquires a response sentence corresponding to the utterance search query from the conversation database 1220 to 1222 and transmits it to the robot 1210. The collection information received by the robot 1210 is stored in the storage 113 or the memory 114 as input collection information 1241.

  On the other hand, the proposal information acquisition unit 1234 acquires proposal information for the user stored in the proposal information database 1240 via the network interface 115 and the network 116. Proposal information is an example of provided information. Each proposal information is, for example, a response sentence including contents to be proposed to the user. The proposal information database 1240 holds a search query associated with each proposal information, and the proposal information acquisition unit 1234 acquires a search query associated with the acquired proposal information from the proposal information database 1240. Hereinafter, a search query associated with each proposal information is referred to as a proposal search query.

  The proposal information acquisition unit 1234 transmits the acquired proposal search query from the network interface 115 to the information collection server 117 via the network 116. The information collection server 117 collects response sentences corresponding to the suggested search query from the conversation databases 1220 to 1222 and returns them to the robot 1210. The robot 1210 stores information including each response sentence corresponding to the proposal search query received from the information collection server 117 as the proposal collection information 1242 in the storage 113 or the memory 114.

  Note that the conversation databases 1220 to 1222 may be included in the storage 113 or the memory 114. At this time, for example, the speech recognition unit 1231 may acquire a response sentence corresponding to the utterance search query, and the proposal information acquisition unit 1234 may acquire a response sentence corresponding to the proposal search query from the conversation databases 1220 to 1222.

  Here, the input collection information 1241 corresponds to the input collection information 141 in the first embodiment, and the proposal collection information 1242 corresponds to the advertisement collection information 142 in the first embodiment. The information presentation system 1200 presents information to the user according to the flowchart of FIG.

  The graph information generation unit 131 generates first graph information from the input collection information 1241 and second graph information from the proposal collection information 1242. The graph integration unit 135 integrates the first graph and the second graph. The route selection unit 136 selects a route from the first vertex indicating the utterance search query to the second vertex indicating the suggested search query. The utterance content generation unit 1233 selects, from the input collection information 1241 and the proposal collection information 1242, proposal related information that is a response sentence including a word corresponding to the vertex on the route selected by the vertex selection unit 137. The utterance content generation unit 1233 generates a response utterance including the selected proposal related information, and outputs the response utterance to the user via the speaker 1218.

  Here, the proposal related information corresponds to the advertisement related information in the first embodiment, the response utterance corresponds to the presentation information in the first embodiment, and the proposal information corresponds to the advertisement information. Further, the utterance content generation unit 1233 generates a response utterance by replacing a predetermined part of a response sentence included in the dialogue scenario 1243 with proposal related information, for example.

  In step S212, the proposal information output unit 1238 determines whether the user has made a response indicating an interest in the proposal related information. For example, when the user gives a positive response to the response utterance, the proposal information output unit 1238 determines that the user has shown interest in the proposal related information. In addition, for example, when the user makes a negative response to the response utterance or does not make a response, the proposal information output unit 1238 determines that the user has not shown interest in the proposal related information. Note that the voice recognition unit 1231 interprets the user's response to determine whether or not the user has made a positive response.

  The robot 1210 may have a physical interface such as a button, for example. At this time, for example, the proposal information output unit 1238 determines that the user is interested in the proposal related information when the user presses the button, and the user is interested in the proposal related information when the user does not press the button. May be determined not to be indicated.

  The robot 1210 outputs the proposal related information to the user according to the route so that the user's interest is gradually guided to the proposal information. When the proposal related information continues to be selected, the utterance about the proposal information is finally given to the user. Output.

  In addition, this invention is not limited to an above-described Example, Various modifications are included. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described. Also, a part of the configuration of a certain embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can be added to the configuration of a certain embodiment. Further, it is possible to add, delete, and replace other configurations for a part of the configuration of each embodiment.

  Each of the above-described configurations, functions, processing units, processing means, and the like may be realized by hardware by designing a part or all of them with, for example, an integrated circuit. Each of the above-described configurations, functions, and the like may be realized by software by interpreting and executing a program that realizes each function by the processor. Information such as programs, tables, and files for realizing each function can be stored in a memory, a recording device such as a hard disk or SSD (Solid State Drive), or a recording medium such as an IC card, SD card, or DVD.

  Further, the control lines and information lines indicate what is considered necessary for the explanation, and not all the control lines and information lines on the product are necessarily shown. Actually, it may be considered that almost all the components are connected to each other.

Claims (15)

An information presentation system for presenting information to a user,
Including a processor and a storage device;
The storage device
A first search query that accepts input from the user;
A first collection information group comprising collection information from a database according to the first search query;
Information provided in advance,
A second search query associated with the provided information;
A second collection information group consisting of collection information from the database by the second search query,
The processor is
Generating a first graph in which the first search query and a word included in the first collection information group are used as vertices, and vertices indicating words included in the same collection information are connected by edges;
Generating a second graph in which the second search query and a word included in the second collection information group are used as vertices, and vertices indicating words included in the same collection information are connected by edges;
Generating an integrated graph obtained by integrating the first graph and the second graph at a common vertex of the first graph and the second graph;
Performing a first process of presenting information to the user;
In the first process,
A vertex on a first path connecting the first vertex indicating the first search query and the second vertex indicating the second search query on the integrated graph, and selecting a vertex other than the first vertex;
When the second vertex is selected, the provision information is presented to the user,
When selecting a vertex other than the second vertex,
Selecting one or more collection information including a word indicating the selected vertex from the first collection information group and the second collection information group;
Presenting at least a portion of the selected one or more collected information to the user;
Information presentation system. The information presentation system according to claim 1,
The processor is
Repeating the first process,
In the first process,
When the second vertex is selected, the provision information is presented to the user, the repetition of the first process is terminated,
When selecting a vertex other than the second vertex and receiving a response indicating an interest in at least a part of the presented vertex from the user, information collected from the database using a word indicating the selected vertex as a search query A group is acquired and stored in the storage device as the first collection information group;
In selecting a vertex on the first path in the first process after the second time, an information presentation system that selects a vertex closer to the second vertex in the first path than the vertex selected in the previous first process . The information presentation system according to claim 1,
The processor is
When the first graph and the second graph do not include a common vertex,
Repeating the expansion process of expanding the first graph or the second graph until the first graph and the second graph include a common vertex,
In the extension process,
A vertex other than the first vertex and the second vertex of one of the first graph and the second graph, which is an unselected vertex in the expansion process, and selects a vertex for expansion,
Acquiring a third collection information group from the database using a word indicating the selected extension vertex as a search query;
Adding a vertex indicating the word included in the third collection information group to the one side,
The information presentation system which connects the said one vertex corresponding to the word contained in the same collection information with an edge. The information presentation system according to claim 3,
In the extension process, the processor selects a vertex other than the one of the first vertex and the second vertex and having a maximum degree centrality from unselected vertices in the extension process. Information presentation system to choose. The information presentation system according to claim 3,
In the expansion process, the processor selects a vertex other than the one of the first vertex and the second vertex and having a maximum intermediary centrality from unselected vertices in the expansion processing. Information presentation system to choose. The information presentation system according to claim 1,
The processor is
Determining the length of each edge of the first graph based on the number of co-occurrence in the first collection information group of words indicating two vertices connected by each edge;
The information presentation system which determines the length of each edge of the said 2nd graph based on the co-occurrence number in the said 2nd collection information group of the word which shows the 2 vertex connected by each edge. The information presentation system according to claim 6,
The processor is
Calculating the distance of each path connecting the first vertex and the second vertex based on the length of the edge included in each path;
An information presentation system that selects the first route based on each of the calculated distances. The information presentation system according to claim 6,
The processor is
For each path connecting the first vertex and the second vertex, the degree of co-occurrence between any two vertices connected by an edge in the path is determined for each of the first graph and the second graph between the two vertices. Based on the length of the edge,
The information presentation system which selects the 1st course from the course where all the computed co-occurrence degrees are more than a threshold. The information presentation system according to claim 1,
The processor includes: a path including a first predetermined number or more of vertices whose degree centrality is equal to or greater than a first threshold; or a path including a second predetermined number or more of vertices whose median centrality is equal to or greater than a second threshold. Choose an information presentation system. The information presentation system according to claim 1,
The information presentation system, wherein in the selection of a vertex on the first route in the first process, the processor selects a vertex closest to the first search query in the first route. The information presentation system according to claim 1,
The processor selects a vertex based on the degree centrality of the vertex on the first path or the median centrality of the vertex on the first path in selecting the vertex on the first path in the first process. Information presentation system. The information presentation system according to claim 1,
The information presentation system, wherein the second search query is a search query input from another user. The information presentation system according to claim 1,
The database is a conversation database indicating correspondence between a search query and a response sentence,
Each collection information included in the first collection information group is a response sentence corresponding to the first search query indicated by the conversation database,
Each of the collection information included in the second collection information group is a response sentence corresponding to the second search query indicated by the conversation database. An information presentation system is a method for presenting information to a user,
The information presentation system includes:
A first search query that accepts input from the user;
A first collection information group comprising collection information from a database according to the first search query;
Information provided in advance,
A second search query associated with the provided information;
A second collection information group consisting of collection information from the database by the second search query,
In the method, the information presentation system includes:
Generating a first graph in which the first search query and a word included in the first collection information group are used as vertices, and vertices indicating words included in the same collection information are connected by edges;
Generating a second graph in which the second search query and a word included in the second collection information group are used as vertices, and vertices indicating words included in the same collection information are connected by edges;
Generating an integrated graph obtained by integrating the first graph and the second graph at a common vertex of the first graph and the second graph;
Performing a first process of presenting information to the user;
In the first process,
A vertex on a first path connecting the first vertex indicating the first search query and the second vertex indicating the second search query on the integrated graph, and selecting a vertex other than the first vertex;
When the second vertex is selected, the provision information is presented to the user,
When selecting a vertex other than the second vertex,
Selecting one or more collection information including a word indicating the selected vertex from the first collection information group and the second collection information group;
Presenting at least a portion of the selected one or more collected information to the user. 15. A method according to claim 14, comprising
The information presentation system includes:
Repeating the first process,
In the first process,
When the second vertex is selected, the provision information is presented to the user, the repetition of the first process is terminated,
When selecting a vertex other than the second vertex and receiving a response indicating an interest in at least a part of the presented vertex from the user, information collected from the database using a word indicating the selected vertex as a search query A group is acquired and stored in the information presentation system as the first collection information group;
A method of selecting a vertex closer to the second vertex in the first route than a vertex selected in the previous first processing in selecting the vertex on the first route in the first processing after the second time.

Images