JP4860439B2 - Automatic question generation system - Google Patents

Description

  The present invention relates to a server, a system, a method, and a program for automatically generating a question sentence. More specifically, a server that creates a question sentence database and sends the question sentence data stored in the question sentence database to the terminal, a question sentence automatic generation system, and a question that is created in the question sentence database and accumulated in the question sentence database The present invention relates to a method for transmitting sentence data to a terminal and a program thereof.

  There are various useful contents on the Internet. In order to obtain the information that the user wants to know, the user naturally searches the Internet using a terminal and obtains information from useful content. And if you have a question or want to know in daily life, search for a keyword from a search site such as a portal site that serves as the entrance to the Internet with a search engine or link collection as the core in order to seek an answer to solve the question. It is often done.

  However, in order for the user to obtain a desired answer, it is necessary to appropriately express the question contents with keywords. This is because an appropriate answer to a question cannot be obtained unless an appropriate keyword is input. However, it is difficult to find out what the appropriate keywords are. Therefore, there are publications related to keyword search.

  In addition, various question sites have been set up by Internet companies and people familiar with the Internet. And the answer to a question is also obtained using the bulletin board of this question site. The bulletin board of the question site corresponds to, for example, “Yahoo! Chiebukuro (http://chiebukuro.yahoo.co.jp/)”. However, on the bulletin board of the question site, even if you fill out the questions, you must look at the bulletin board and wait for a third party with knowledge about the question to answer, You cannot get an answer to. Also, even if a third party answers, if the number of answers is small, it may be anxious whether the answers really match.

  In order to obtain an appropriate answer to the question as much as possible, for example, the following various methods are considered. First, as a method for obtaining an answer to a question, a pair of a question and an answer corresponding to the inputted question is searched, and whether or not the pair of the question and the answer matches the inputted question A method of performing evaluation and displaying preferentially those having a high evaluation value is disclosed (for example, Patent Document 1).

  Secondly, the question text is parsed to extract keywords, the question type is identified by information search and the field is identified by question type detection, and the result with higher accuracy is added to the result with lower accuracy. A method of providing feedback and outputting answer candidates is disclosed (for example, Patent Document 2).

Thirdly, a method is disclosed in which a correspondence between a question article and an answer article is taken, a group of those articles is stored as a database, and an answer to a user's question is searched from the database and presented (for example, Non-Patent Document 1).
JP 2003-006207 A JP 2006-119697 A Satoshi Yokota, Hiroyuki Sakai, Shigeru Masuyama, Extraction of question articles from a Web bulletin board for a question answering system, Proc. Of the 12th Annual Conference of the Association for Natural Language Processing (NLP2006), p. 332-335

  The method described in the above-mentioned document extracts and presents the answer to the inputted question sentence by various methods. In addition, since these are premised on the input of a question sentence, when an answer is obtained by these methods on the premise of inputting one or a plurality of keywords, the result must be low accuracy. .

  Therefore, the present invention can solve the above-mentioned problems, create a question sentence database in advance, and automatically return a question sentence suitable for the user's intention from the created question sentence database to the user terminal. The purpose is to provide a server.

  For the above purpose, the following are specifically provided.

(1) A server that creates a question sentence database and transmits question sentence data accumulated in the question sentence database to a user terminal connected via a network,
Means for accumulating content data from a content server connected to the network according to data of a URL list set in advance;
Means for extracting question sentence data that matches the extraction condition from the accumulated content data in accordance with a preset extraction condition;
Means for storing the extracted question sentence data in a question sentence database;
Means for receiving text data transmitted from the user terminal;
Means for retrieving first question text data related to the text data using the question text database based on the received text data;
Means for transmitting the retrieved first question sentence data to the user terminal,
The means for receiving further receives the first question sentence data transmitted to the user terminal,
The server that searches the second question sentence data further related to the first question sentence data using the question sentence database based on the received first question sentence data.

  According to the configuration of (1), the server accumulates content data from the content server connected to the network according to the URL list data set in advance, and stores the accumulated content data according to the preset extraction condition. Question text data that matches the extraction conditions is extracted from the list. The extracted question sentence data can be stored in the question sentence database. Here, the content server is any server on the network. Further, the server receives text data that is a question sentence transmitted from a user terminal connected to the network, and based on the received text data, first question sentence data related to the text data using a question sentence database And the searched first question sentence data can be transmitted to the user terminal. Here, the text data refers to, for example, a question sentence input from the user terminal, and corresponds to a keyword or a question document. Further, the server receives the first question sentence data transmitted to the user terminal, and based on the received first question sentence data, the second question related to the first question sentence data using the question sentence database. Sentence data can be searched.

  As a result, the question text data extracted from the content data and stored in the question text database of the server is searched based on the text data that is the question text such as the keyword transmitted from the user terminal, and the question related to the text data is obtained. It can be sent to the user terminal as sentence data. Therefore, even when the text data input from the user terminal cannot properly indicate the question contents, the question sentence data related to the text data can be transmitted to the user terminal, and the user's intention You can return the question text that was there.

  In particular, it is difficult for the user to create a question sentence when the subject to be examined is vague or the specific content to be examined is specialized. Even in such a case, a question sentence suitable for the user's intention can be found based on the keyword.

  Moreover, it is possible to repeat a plurality of times to obtain further related question sentence data based on such question sentence data. Therefore, it is possible to obtain highly accurate question sentence data that meets the user's intention.

  (2) The server according to (1), wherein the URL list data set in advance is data specifying a bulletin board for knowledge search among the content data of the content server.

  According to the configuration of (2), the URL list data set in advance can be data specifying a bulletin board for knowledge search among the content data of the content server.

  Thus, information can be obtained from useful content by creating question sentence data to be stored in the question sentence database based on data posted on a bulletin board for knowledge search. Therefore, a highly accurate question sentence database can be created. The URL list data can be set by the administrator. Therefore, even when a new bulletin board for knowledge search is added or deleted, it can be easily performed.

  (3) The server according to (1) or (2), wherein the preset extraction condition uses a heuristic rule using an empirical rule.

  According to the configuration of (3), a heuristic rule using an empirical rule can be used as the preset extraction condition.

  Thus, the question sentence data stored in the question sentence database can be classified as a question sentence by a heuristic technique found as an empirical rule. Therefore, a highly accurate question sentence database can be created. Heuristic rules can be set by the administrator. Therefore, new rules can be easily reflected.

  (4) The searching means performs morphological analysis of the text data or the first question sentence data, and the question sentence data stored in the question sentence database for each result data divided by the morphological analysis. TF * IDF is calculated from the total number of the question sentence data stored in the question sentence database, the appearance frequency, and the number of question sentence data. The server according to any one of (1) to (3), which is obtained and performed based on the numerical data of the TF * IDF.

  According to the configuration of (4), the server performs morphological analysis of the text data or the first question sentence data, and in the question sentence data stored in the question sentence database for each result data divided by the morphological analysis. The appearance frequency (number of corresponding data) and the number of question sentence data (number of question sentences including the corresponding data) are calculated. Further, the server calculates TF * IDF (a kind of machine learning) from the total number of question sentence data stored in the question sentence database, the appearance frequency in the question sentence data obtained by calculation, and the number of question sentence data. ) And a search can be performed based on the calculated numerical data of TF * IDF.

  This makes it possible to search using the machine learning function as described above. Therefore, a highly accurate search process using the question sentence database can be executed. In addition, since all are performed mechanically, high-speed processing can be performed.

(5) means for accumulating content data from a content server connected to the network according to URL list data set in advance;
Means for extracting question sentence data that matches the extraction condition from the accumulated content data in accordance with a preset extraction condition;
The server according to any one of (1) to (4), wherein control is performed so that the means for storing the extracted question sentence data in a question sentence database is executed at a predetermined timing.

  According to the configuration of (5), the server accumulates the content data from the content server connected to the network according to the data of the URL list set in advance, and stores the accumulated content data according to the preset extraction condition. It is possible to perform control so that extraction of the question sentence data that matches the extraction condition from the inside and storing the extracted question sentence data in the question sentence database is performed at a predetermined timing. The predetermined timing refers to, for example, a timing that is carried out once a day or every weekend.

  Thus, the question sentence database can be created in advance and periodically updated at a predetermined timing. Therefore, since new useful information can be taken into the question sentence database at any time, a highly accurate question sentence database can be automatically created.

(6) means for obtaining answer data for the question sentence data based on the question sentence data;
The server according to any one of (1) to (5), further comprising: means for transmitting the acquired answer data together with the question sentence data to the user terminal.

  According to the configuration of (6), the server can acquire answer data for the question sentence data based on the question sentence data, and transmit the acquired answer data to the user terminal together with the question sentence data.

  Thus, an answer can be obtained based on a question sentence that matches the user's intention. That is, after generating appropriate question sentence data, an answer based on the question sentence data can be obtained. Therefore, the user can obtain an appropriate answer to the question.

(7) A question sentence in which a user terminal and a server that creates a question sentence database and transmits the question sentence data stored in the question sentence database to the user terminal are connected to each other via a network. An automatic generation system,
The server
Means for accumulating content data from a content server connected to the network according to data of a URL list set in advance;
Means for extracting question sentence data that matches the extraction condition from the accumulated content data in accordance with a preset extraction condition;
Means for storing the extracted question sentence data in a question sentence database;
Means for receiving text data transmitted from the user terminal;
Means for retrieving first question text data related to the text data using the question text database based on the received text data;
Means for transmitting the retrieved first question sentence data to the user terminal,
The user terminal is
Means for receiving the first question sentence data transmitted from the server;
Means for displaying the received first question sentence data;
Means for selecting the displayed first question sentence data;
Means for transmitting the selected first question sentence data,
The server
Means for receiving the first question sentence data transmitted by the user terminal;
Means for searching for second question text data related to the first question text data using the question text database based on the received first question text data; system.

  According to the structure of (7), the system which has the effect | action and effect similar to said (1) can be provided.

(8) The server
A keyword search using the content data of the content server based on the text data or the first question sentence data, and means for transmitting result data to the user terminal;
The receiving means included in the user terminal further receives the result data,
The question sentence automatic generation system according to (7), wherein the display unit included in the user terminal further displays the result data.

  According to the configuration of (8), the server can perform keyword search using the content data of the content server based on the text data or the first question sentence data, and transmit the result data to the user terminal. Further, the user terminal can receive the keyword search result data in addition to the search result using the question sentence database, and can display both of the received search result data.

  This allows the user terminal to perform keyword search based on question sentence data such as text data or related question sentence data together with result data obtained by searching the question sentence database based on text data that is a question sentence or related question sentence data. As a result, the result data can be displayed. Therefore, for the user, both the result data obtained by simply performing a keyword search on the input data and the result data obtained by searching the question sentence database after analyzing the input data are displayed at a time. Convenience will be improved.

(9) A server connected to a network creates a question sentence database and transmits the question sentence data stored in the question sentence database to a user terminal connected to the network,
Accumulating content data from a content server connected to the network according to URL list data set in advance;
Extracting the question sentence data that matches the extraction condition from the accumulated content data according to a preset extraction condition;
Storing the extracted question sentence data in a question sentence database;
Receiving text data transmitted from the user terminal;
Searching for first question text data related to the text data using the question text database based on the received text data;
Transmitting the retrieved first question sentence data to the user terminal;
Receiving the first question sentence data transmitted to the user terminal;
Searching for second question text data related to the first question text data using the question text database based on the received first question text data.

  According to the structure of (9), the method which has the same effect | action and effect as said (1) can be provided.

(10) A program for causing a server connected to a network to create a question sentence database and transmitting question sentence data stored in the question sentence database to a user terminal connected to the network,
Accumulating content data from a content server connected to the network according to URL list data set in advance;
Extracting the question sentence data that matches the extraction condition from the accumulated content data according to a preset extraction condition;
Storing the extracted question sentence data in a question sentence database;
Receiving text data transmitted from the user terminal;
Searching for first question text data related to the text data using the question text database based on the received text data;
Transmitting the retrieved first question sentence data to the user terminal;
Receiving the first question sentence data transmitted to the user terminal;
A program for causing the server to execute, based on the received first question sentence data, a step of searching for second question sentence data related to the first question sentence data using the question sentence database.

  According to the configuration of (10), it is possible to provide a program having the same operation and effect as the above (1).

  According to the present invention, even when the text data that is the question text input from the user terminal cannot properly indicate the question content, the question text data related to the text data is transmitted to the user terminal. By repeating this, it is possible to provide a server that can create a question sentence database in advance, automatically extract a question sentence suitable for the user's intention from the created question sentence database, and return it to the user terminal.

  Hereinafter, the best mode for carrying out the present invention will be described with reference to the drawings. This is merely an example, and the technical scope of the present invention is not limited to this.

[System overall configuration]
FIG. 1 is a block diagram showing an overall configuration of a computer system 1 and a functional configuration of a server 10 according to an example of a preferred embodiment of the present invention.

  Server 10 and user terminal 20 are connected via communication line 30. Here, there is no restriction | limiting in the number of the servers 10, You may comprise by 1 or multiple as needed. Further, the server 10 may be configured to include a Web server, a DB server, and an application server as necessary, and may be configured with one server or different servers.

  The communication line 30 indicates the Internet, for example. The communication line 30 is connected to a content server 15 having a content DB 65 in which various contents are stored. There are a large number of content servers 15 all over the world.

  The server 10 functions as at least a question sentence automatic generation device, and the storage unit 60 includes at least a question sentence DB 61.

  Further, the control unit 50 controls the entire server 10, and at least the content accumulation means 51, the question sentence data extraction means 52, the question sentence data storage means 53, the text data reception means 54, the question sentence data search means 55, and the result data A transmission means 56 is provided.

  The server 10 accumulates content data from the content DB 65 of the content server 15 connected via the communication line 30 in accordance with URL list data set in advance. Further, according to preset extraction conditions, question sentence data matching the extraction conditions is extracted from the accumulated content data, and the extracted question sentence data is stored in the question sentence DB 61. Furthermore, the server 10 receives text data which is a question sentence such as a keyword transmitted from the user terminal 20, and searches for question sentence data related to the text data using the question sentence DB 61 based on the received text data. Then, the retrieved question sentence data is transmitted to the user terminal 20.

[Conceptual diagram of processing]
2 and 3 are conceptual diagrams of processing according to an example of a preferred embodiment of the present invention. FIG. 2 is a conceptual diagram illustrating a question sentence DB 61 creation process performed by the server 10 which is one of processes according to an example of the preferred embodiment of the present invention.

  The server 10 has a URL list related to content accumulation in the storage unit 60 in advance. For example, as shown in the upper left diagram of FIG. 2, the URL list related to content accumulation is a question and answer such as “XX! Wisdom bag”, “Tell me! △△△”, “XXX bulletin board”. This corresponds to the URL list of the site on which is posted.

  The crawler, which is the content accumulation means 51 of the server 10, refers to the URL list related to content accumulation ((1) in FIG. 2) and makes a URL request to the content server 15 on the Internet. As a result, the content of the corresponding URL can be received ((2) in FIG. 2). The crawler refers to a search robot such as a Web patrol program, and periodically acquires documents, images, and the like on the Web.

  Thereafter, the server 10 extracts a question sentence from the accumulated content by the question sentence data extraction unit 52 ((3) in FIG. 2). Heuristic rules using heuristics are used to extract question sentences. The heuristic rule means “a heuristic method” or “a heuristic solution”, and does not necessarily lead to a correct answer for a certain problem, but indicates a solution that obtains a solution as close as possible. In the present invention, when extracting a question sentence, it is performed using the character “question” or “question”, which is the characteristic of the question sentence found from the rule of thumb. When these characters are found, a sentence including these characters is extracted as a question sentence.

  Thereafter, the question sentence data storage unit 53 of the server 10 stores only the extracted question sentence data in the question sentence DB 61 ((4) in FIG. 2).

  The effect of the present invention can be found by using the question sentence DB 61 created by the process of creating the question sentence DB 61 for the question sentence automatic generation process described later.

  FIG. 3 is a conceptual diagram about the question sentence automatic generation process of the server 10 which is one of processes according to an example of the preferred embodiment of the present invention.

  When a user inputs a question text in a question input text box displayed on the user terminal 20, question text data is transmitted to the server 10 ((1) in FIG. 3). The text data receiving means 54 of the server 10 receives the question sentence data, and uses the question sentence data, the question sentence data related to the question sentence data by the search engine which is the question sentence data search means 55 of the server 10 Is retrieved from the question sentence DB 61 ((2) to (4) in FIG. 3). Thereafter, the result data transmission means 56 of the server 10 transmits the related question sentence data, which is the data of the search result, to the user terminal 20 ((5) in FIG. 3). The user terminal 20 can display the received related question sentence data.

  When it is desired to obtain further related question text data from the related question text data, the related question text data is transmitted again to the server 10 ((6) in FIG. 3). Hereinafter, the server 10 performs a search in the same manner as the above-described processing based on the related question sentence data, and transmits the related question sentence data that is data of the search result to the user terminal 20. Thereby, the user terminal 20 can display the received related question sentence data.

[Question sentence DB 61 creation process flow]
FIG. 4 is a flowchart showing the flow of the question sentence DB 61 creation process of the server 10 according to an example of the preferred embodiment of the present invention.

  First, in step S1, the control unit 50 of the server 10 determines whether or not a predetermined time has elapsed. When the predetermined time has elapsed (when YES is determined in the process of step S1), the control unit 50 of the server 10 moves the process to step S2. On the other hand, if the predetermined time has not elapsed (NO is determined in step S1), the control unit 50 of the server 10 continues to stay in step S1 and waits for the predetermined time to elapse. .

  In step S <b> 2, the content accumulation unit 51 of the server 10 transmits a URL request based on a URL list stored in advance in the storage unit 60 of the server 10. Specifically, after a URL list stored in advance in the storage unit 60 of the server 10 is sequentially read from the beginning, a request for content data is transmitted to the corresponding URL. Then, the control part 50 of the server 10 moves a process to step S3.

  In step S3, the control unit 50 of the server 10 receives content data transmitted from the content server 15 connected to the communication line 30 based on the URL request. Then, the control part 50 of the server 10 moves a process to step S4.

  In step S4, the question sentence data extracting unit 52 of the server 10 uses the question sentence data from the content data received in step S3 based on the heuristic rule (empirical rule) stored in the storage unit 60 of the server 10 in advance. To extract. Specifically, for each content data, whether the heuristic rules stored in advance in the storage unit 60 of the server 10 are read in order from the beginning, and the text stored as the heuristic rules is described in the content data. Confirm whether or not. If there is a written word, one sentence including the word is extracted. This process is performed for all the content data received in step S3. Thereafter, the control unit 50 of the server 10 moves the process to step S5.

  In step S5, the question sentence data storage unit 53 of the server 10 stores the question sentence data extracted in step S4 in the question sentence DB 61 of the storage unit 60. Thereafter, the control unit 50 of the server 10 ends this process.

  As described in the present embodiment, it is desirable that the process for creating the question sentence DB 61 is periodically performed at a predetermined timing. The predetermined timing can be realized by using, for example, a schedule function of the server 10, and indicates that the predetermined timing is performed at regular intervals such as every day or every weekend. This is because the content stored in the content DB 65 of the content server 15 connected to the communication line 30 is updated every day, and new information increases every moment. Therefore, by periodically performing this process, it is possible to obtain question sentence data that matches the times at an early stage. In addition, you may perform with a search process triggered by the process which searches the related question sentence mentioned later.

[Question answer processing flow using question sentence automatic generation processing]
5 to 7 are flowcharts showing the flow of the question answering process using the question sentence automatic generation process of the server 10 according to an example of the preferred embodiment of the present invention.

  First, FIG. 5 is a flowchart showing a flow of processing for the server 10 to transmit related question message data when a question message is input to the user terminal 20.

  First, the process of the user terminal 20 will be described. In step S201, the control unit of the user terminal 20 displays a question sentence input screen. This question sentence input screen shows, for example, the one shown in FIG. Thereafter, when a question text is input by the user and a transmission instruction is accepted, the control unit of the user terminal 20 transmits text data that is the input question text to the server 10 in step S202.

  When the control unit of the user terminal 20 receives the related question message data transmitted from the server 10 (step S203), in step S204, the control unit of the user terminal 20 displays a related question message list. This related question sentence list shows, for example, a related question sentence search result 73 of (2) in FIG. Thereafter, the process proceeds to step S205 in FIG. 6 or step S208 in FIG. In addition, you may complete | finish this process after outputting related question sentence data.

  Next, processing of the server 10 will be described. In response to the transmission of text data from the user terminal 20, in step S101, the text data receiving unit 54 of the server 10 receives the text data. Thereafter, the text data receiving means 54 of the server 10 moves the process to step S102.

  In step S102, the question sentence data search means 55 of the server 10 performs a related question sentence search process. Details of this search processing will be described later with reference to FIG. Thereafter, the question sentence data search means 55 of the server 10 moves the process to step S103.

  In step S <b> 103, the result data transmission unit 56 of the server 10 transmits the related question text data searched by the related question text search processing to the user terminal 20. Then, according to the process of the user terminal 20, the process of the server 10 is moved to step S104 in FIG. 6 or step S107 in FIG. If request data is not transmitted from the user terminal 20, the control unit 50 of the server 10 ends this process.

  Next, FIG. 6 is a flowchart showing a flow of processing for the server 10 to further transmit related question message data when the user terminal 20 receives the related question message data transmitted from the server 10. .

  First, processing of the user terminal 20 will be described. In step S205, the control unit of the user terminal 20 re-searches in response to the user selecting one question sentence data from the related question sentence list displayed in step S204 of FIG. 5 and instructing the re-search. The question sentence data selected by the user as data is transmitted to the server 10.

  When the control unit of the user terminal 20 receives the related question message data transmitted from the server 10 (step S206), in step S207, the control unit of the user terminal 20 displays a related question message list. Thereafter, the process proceeds to step S208 in FIG. In addition, you may complete | finish this process after outputting related question sentence data.

  Next, processing of the server 10 will be described. In response to the transmission of one question sentence data from the user terminal 20, in step S104, the text data receiving unit 54 of the server 10 receives one question sentence data as text data. Thereafter, the text data receiving means 54 of the server 10 moves the process to step S105.

  In step S105, the question sentence data search means 55 of the server 10 performs a related question sentence search process. Details of this search processing will be described later with reference to FIG. Thereafter, the question sentence data search means 55 of the server 10 moves the process to step S106.

  In step S <b> 106, the result data transmission unit 56 of the server 10 transmits the related question text data searched by the related question text search processing to the user terminal 20. Then, according to the process of the user terminal 20, the process of the server 10 is moved to step S107 of FIG. If request data is not transmitted from the user terminal 20, the control unit 50 of the server 10 ends this process.

  Finally, FIG. 7 shows that when the related question message data transmitted from the server 10 is input to the user terminal 20, the server 10 transmits search result data which is answer data corresponding to the question message data. It is a flowchart which shows the flow of a process of.

  First, processing of the user terminal 20 will be described. In step S208, the control unit of the user terminal 20 responds that the user selects one question sentence data from the related question sentence list displayed in step S204 of FIG. 5 or step S207 of FIG. Then, the response request data of the related question sentence list is transmitted to the server 10.

  When the control unit of the user terminal 20 receives search result data that is response data to the related question sentence data transmitted from the server 10 (step S209), in step S210, the control unit of the user terminal 20 displays an answer result screen. indicate. Thereafter, the control unit of the user terminal 20 ends this process.

  Next, processing of the server 10 will be described. In response to the transmission of the response request data for the related question list from the user terminal 20, in step S107, the control unit 50 of the server 10 receives the response request data. Thereafter, the control unit 50 of the server 10 moves the process to step S108.

  In step S108, the control unit 50 of the server 10 performs content search processing based on the response request data received in step S107. Specifically, the content server 15 connected to the communication line 30 is searched using the related question text data as a key. Then, those in which answers to the related question sentence data to be searched are described are extracted. Thereafter, the control unit 50 of the server 10 moves the process to step S109.

  In step S109, the control unit 50 of the server 10 transmits the search result data extracted in step S108 to the user terminal 20. Thereafter, the control unit 50 of the server 10 ends this process.

[Related question text search processing flow]
FIG. 8 is a flowchart showing the flow of related question text search processing of the server 10 according to an example of the preferred embodiment of the present invention.

  First, in step S11, the question sentence data search means 55 of the server 10 performs morphological analysis of search target data. Morphological analysis refers to dividing a sentence written in a natural language into minimum unit data that has meaning in the language of morpheme. Thereafter, the question sentence data search means 55 of the server 10 moves the process to step S12.

  In step S12, the question sentence data search means 55 of the server 10 puts the result data obtained by dividing the search target data in step S11 into the stack of the storage unit 60. Thereafter, the question sentence data search means 55 of the server 10 moves the process to step S13.

その後、サーバ１０の質問文データ検索手段５５は、処理をステップＳ１４に移す。 In step S13, the question sentence data search means 55 of the server 10 extracts one piece of data from the stack of the storage unit 60, and calculates a TF * IDF value using the question sentence DB 61. The TF * IDF value is calculated based on the appearance frequency of corresponding data among a plurality of document data to be searched, the number of document data including the data, and the total number of document data to be searched. It is a scale that gives higher scores to data that appears biased to specific document data. It is calculated by the following formula.

Thereafter, the question sentence data search means 55 of the server 10 moves the process to step S14.

  In step S14, the question sentence data search means 55 of the server 10 determines whether or not the stack of the storage unit 60 is empty. When the stack is empty (when YES is determined in the process of step S14), the question sentence data search unit 55 of the server 10 moves the process to step S15. On the other hand, when the stack is not empty (when NO is determined in the process of step S14), the question sentence data search means 55 of the server 10 moves the process to step S13.

  In step S15, the question sentence data search means 55 of the server 10 ranks based on the TF * IDF value calculated in step S14. The ranking is given so that the higher the score of the TF * IDF value, the higher the ranking. Thereafter, the question sentence data search means 55 of the server 10 moves the process to step S16.

  In step S16, the question sentence data search means 55 of the server 10 performs an OR search on the data in the question sentence DB 61 based on the top five ranking data assigned in step S14. Thereafter, the question sentence data search means 55 of the server 10 moves the process to step S17.

  In step S17, the question sentence data search means 55 of the server 10 acquires the search result data searched in step S16, and ends this process.

  In step S16, the OR search is performed on the data in the question sentence DB 61 based on the top five data of the ranking given in step S14. However, the present invention is not limited to this. For example, an OR search may be performed on the upper four data or the upper six data.

[Screen Description]
FIG. 9 shows an example of a screen displayed on the user terminal 20 according to an example of the preferred embodiment of the present invention.

  First, a screen on which a text box 74 is displayed as the question input 71 is displayed on the user terminal 20 ((1) in FIG. 9). In this text box 74, the user inputs the question content as a text document. Thereafter, when the transmission button 75 is clicked, a processing result screen is displayed on the user terminal 20 after the processing in the server 10.

  The processing result screen is a screen including a keyword search result 72 and a related question sentence search result 73 in addition to the question input 71 ((2) in FIG. 9). Here, the question input 71 is input by the user into the text box 74 when there is a new question, as in (1). Next, the keyword search result 72 is obtained by performing a keyword search based on the text document input in the text box 74 of (1) and displaying the result data. The result data is data received based on the search result from the content DB 65 of the content server 15 connected via the communication line 30. In this example, a URL link is provided as a search result in the keyword search result display unit 76, and when the URL link is clicked, the corresponding page is displayed as a new screen ((3) in FIG. 9). . The keyword search result display section 76 may display contents other than the URL link, or may display the first few lines of the search result contents as a digest.

  Finally, the related question sentence search result 73 is a result of a related question sentence search using the question sentence DB 61 based on the text document input in the text box 74 of (1) and displaying the result. The result data is data received based on the search results from the question sentence DB 61 stored in the storage unit 60 of the server 10. When one data is selected from the data of the related question sentence search result display unit 77 displayed as the result data and the re-search button 79 is clicked, the related question sentence search process is performed again based on the selected one data. The result is similarly displayed as shown in (2) of FIG. Further, when one data is selected from the related question sentence search result display section 77 data displayed as the result data and the display button 78 is clicked, the answer data is searched for the answer data for the selected one data. Is displayed as a new screen ((4) in FIG. 9).

[Example 1]
FIG. 10 shows a first example of the related question sentence search process according to an example of the preferred embodiment of the present invention.

  First, when a question “data cannot be read” is input from the user terminal 20 ((1) in FIG. 10), the control unit 50 of the server 10 searches using the question sentence DB 61 of the storage unit 60 ( (2) in FIG. 10, as related question sentence search results 73, for example, “1. There is an unprecedented sound from the hard disk,“ ... ”, and“ 2. IH cooking heater is searched. In Europe, there are both opinions of “popular” and “not widespread”… ”,“ 3. I tried to shoot a DVD-R with a DVD video camera and edit it with ○ △ □ But ... " In this way, a question sentence candidate as an event in which data cannot be read can be displayed for a question “data cannot be read”. All of the information obtained as the search results is stored in the question sentence DB 61.

[Example 2]
FIG. 11 shows a second example of related question sentence search processing according to an example of the preferred embodiment of the present invention.

  In the case where "question" and "unprecedented ... from hard disk" as one question sentence data is selected from the related question sentence search results displayed as the related question sentence search result 73 in FIG. The control unit 50 of the server 10 performs morphological analysis on this document ((1) in FIG. 11). This document is divided into data such as “hard disk”, “sound”, “data”,... By morphological analysis. In this example, nouns are extracted as divided data. Based on the morphological analysis of each piece of data ((2) in FIG. 11), ranking is performed after calculating the TF * IDF value, and an OR search is performed by sequentially arranging the top words in the ranking (FIG. 11). (3)). For example, an OR search using the upper five words is performed. The result is displayed again on the user terminal 20 as the related question text search result 73 ((4) in FIG. 11).

  As shown in Fig. 10, when you search for the question "I can't read data", questions about IH cooking heaters and DVD camcorders are also displayed. However, as shown in Fig. 11, information related to personal computer-related data cannot be read. By searching again, using morphological analysis and using TF * IDF, the data related to the personal computer is narrowed down to questions that cannot be read. Therefore, a question sentence close to the question sentence required by the user can be displayed.

[Hardware Configuration of Server 10]
FIG. 12 is a diagram illustrating a hardware configuration of the server 10 according to an example of the preferred embodiment of the present invention. The server 10 constitutes a control unit 50 including a content accumulation unit 51, a question sentence data extraction unit 52, a question sentence data storage unit 53, a text data reception unit 54, a question sentence data search unit 55, and a result data transmission unit 56. CPU (Central Processing Unit) 110 (a plurality of CPUs such as CPU 120 may be added in a multiprocessor configuration), bus line 105, communication I / F 140, main memory 150, BIOS (Basic Input Output System) 160, USB port 190 , I / O controller 170, input means such as keyboard and mouse 180, and display device 122.

  Storage means such as a tape drive 172, a hard disk 174, an optical disk drive 176, and a semiconductor memory 178 can be connected to the I / O controller 170.

  The BIOS 160 stores a boot program executed by the CPU 110 when the server 10 is started up, a program depending on the hardware of the server 10, and the like.

  The hard disk 174 stores various programs for functioning as the server 10 and programs for executing the functions of the present invention, and can configure various databases as necessary.

  As the optical disk drive 176, for example, a DVD-ROM drive, a CD-ROM drive, a DVD-RAM drive, or a CD-RAM drive can be used. In this case, the optical disk 177 corresponding to each drive is used. A program or data may be read from the optical disk 177 by the optical disk drive 176 and provided to the main memory 150 or the hard disk 174 via the I / O controller 170. Similarly, the tape medium 171 corresponding to the tape drive 172 can be used mainly for backup.

  The program provided to the server 10 is provided by being stored in a recording medium such as the hard disk 174, the optical disk 177, or a memory card. The program may be installed in the server 10 and executed by being read from the recording medium via the I / O controller 170 or downloaded via the communication I / F 140.

  The above-described program may be stored in an internal or external storage medium. Here, in addition to the hard disk 174, the optical disk 177, or the memory card, a magneto-optical recording medium such as an MD, or a tape medium 171 can be used as a storage medium constituting the storage unit 60 including the question sentence DB 61. Further, a storage device such as a hard disk 174 or an optical disk library provided in a server system connected to a communication line 30 such as a dedicated communication line or the Internet is used as a recording medium, and a program is provided to the server 10 via the communication line 30. May be.

  Here, the display device 122 displays a screen for accepting input of data to the administrator of the server 10 or displays a screen of a result of calculation processing by the server 10. The display device 122 is a cathode ray tube display (CRT), liquid crystal display, or the like. A display device such as a display device (LCD) is included.

  Here, the input means is for accepting input by the administrator of the server 10 and may be constituted by a keyboard and a mouse 180 or the like.

  The communication I / F 140 is a network adapter that enables the server 10 to be connected to a terminal via a dedicated network or a public network. The communication I / F 140 may include a modem, a cable modem, and an Ethernet (registered trademark) adapter.

  In the above example, the server 10 is mainly described. However, the functions described above can be realized by installing a program in a computer and causing the computer to operate as the server 10. Therefore, the functions realized by the server 10 described as one embodiment in the present invention are realized by executing the above-described method by the computer, or by introducing the above-mentioned program into the computer and executing it. Is possible.

  As mentioned above, although embodiment of this invention was described, this invention is not restricted to embodiment mentioned above. The effects described in the embodiments of the present invention are only the most preferable effects resulting from the present invention, and the effects of the present invention are limited to those described in the embodiments of the present invention. is not.

1 is a block diagram illustrating an overall configuration of a computer system and a functional configuration of a server according to an example of a preferred embodiment of the present invention. It is a conceptual diagram about the preparation process of question sentence DB which is one of the processes which concern on this invention. It is a conceptual diagram about the question sentence automatic generation process which is one of the processes which concern on this invention. It is a flowchart which shows the question sentence DB creation process which concerns on this invention. It is a flowchart which shows the question response process using the question sentence automatic generation process which concerns on this invention. It is a flowchart following FIG. It is a flowchart following FIG. 5 or FIG. It is a flowchart which shows the related question sentence search process which concerns on this invention. It is a figure which shows the example of the screen displayed on the user terminal which concerns on this invention. The Example of the related question sentence search which concerns on this invention is shown. The Example of the related question sentence search which concerns on this invention is shown. It is a figure which shows an example of the hardware constitutions of the server which concerns on this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Computer system 10 Server 15 Content server 20 User terminal 30 Communication line 50 Control part 51 Content accumulation means 52 Question sentence data extraction means 53 Question sentence data storage means 54 Text data reception means 55 Question sentence data search means 56 Result data transmission means 60 storage unit 61 question sentence DB
65 Content DB

Claims (8)

A server for creating a question sentence database and transmitting question sentence data stored in the question sentence database to a user terminal connected via a network,
Means for accumulating content data from a content server connected to the network according to data of a URL list set in advance;
Means for extracting question sentence data that matches the extraction condition from the accumulated content data in accordance with a preset extraction condition;
Means for storing the extracted question sentence data in a question sentence database;
In response to receiving a search request for text data from the user terminal, a plurality of first question sentence data related to the text data for all the question sentence data stored in the question sentence database. Question sentence data search means for searching ;
A result data transmission unit that transmits a page including the list of the plurality of first question sentence data searched and a selection unit to the user terminal;
The question data retrieval means, said when further receiving a search request one of the first question data selected by said selecting means to the user terminal from Oite said plurality of first question data, the searching a plurality of second question data further related to the first question data of the one received by for all question data stored in the question database,
The result data transmission means transmits a page including a list of the plurality of second question sentence data searched and a selection means to the user terminal,
When the user terminal further receives a search request for one second question sentence data selected by the selection means from the plurality of second question sentence data, the one second question sentence data is As the first question sentence data, the question sentence data search means and the result data transmission means are repeatedly executed,
Performing control so that the means for collecting, the means for extracting, and the means for storing are executed at a predetermined timing;
Means for acquiring response data for the received question sentence data from the content server when receiving an answer request for the first question sentence data or the first second question sentence data from the user terminal; ,
Means for transmitting the obtained answer data together with the received question text data to the user terminal;
Ru equipped with a server. The data preset URL list is data specifying the bulletin board for the knowledge search of the content data which the content server has server of claim 1. The server according to claim 1, wherein the preset extraction condition uses a heuristic rule using an empirical rule. The question data retrieval means performs morphological analysis of the first question data of the text data or the one for each result divided data by the morphological analysis, the question message is stored in the question database TF * IDF is calculated from the total number of the question sentence data stored in the question sentence database, the appearance frequency, and the number of question sentence data. the calculated, performed on the basis of the numerical data of the TF * IDF, as claimed in any one of claims 1 to 3 server. A question sentence automatic generation system in which a user terminal and a server that creates a question sentence database and transmits question sentence data stored in the question sentence database to the user terminal are connected to each other via a network Because
The server
Means for accumulating content data from a content server connected to the network according to data of a URL list set in advance;
Means for extracting question sentence data that matches the extraction condition from the accumulated content data in accordance with a preset extraction condition;
Means for storing the extracted question sentence data in a question sentence database;
Wherein Depending user terminal or Latte text particular has received a search request for data, a plurality of first question associated with the text data for all question data stored in the question database Question sentence data retrieval means for retrieving data ;
A result data transmission unit that transmits a page including the list of the plurality of first question sentence data searched and a selection unit to the user terminal;
The server performs control so that the means for collecting, the means for extracting, and the means for storing are executed at a predetermined timing,
The user terminal is
Means for receiving a page including a list of the plurality of first question sentence data transmitted from the server and the selection means ;
Means for displaying the received page ;
And means for transmitting the search request one of the first question data selected by said selecting means from the plurality of first question data included in the displayed the page,
The question data retrieval means of the server, when further receiving a search request for the first question data of the one, the received intended for all question data stored in the question database A plurality of second question sentence data further related to the one first question sentence data ;
The result data transmission unit of the server transmits a page including the list of the plurality of second question sentence data searched and a selection unit to the user terminal,
The means for receiving of the user terminal receives a page including a list of the plurality of second question sentence data transmitted from the server and the selection means,
The means for displaying of the user terminal displays the received page,
When the server further receives a search request for one second question sentence data selected by the selection means from the plurality of second question sentence data in the user terminal, the first second Using the question sentence data as the first question sentence data, repeatedly executing the question sentence data search means and the result data transmission means,
Means for acquiring response data for the received question sentence data from the content server when receiving an answer request for the first question sentence data or the first second question sentence data from the user terminal; ,
Means for transmitting the obtained answer data to the user terminal together with the received question sentence data . The server
Further comprising means for performing a keyword search using the content data of the content server based on the text data or the one first question sentence data, and transmitting result data to the user terminal,
The receiving means included in the user terminal further receives the result data,
The question sentence automatic generation system according to claim 5 , wherein the display unit included in the user terminal further displays the result data. A server connected to a network creates a question sentence database and transmits the question sentence data stored in the question sentence database to a user terminal connected to the network,
Accumulating content data from a content server connected to the network according to URL list data set in advance;
Extracting the question sentence data that matches the extraction condition from the accumulated content data according to a preset extraction condition;
Storing the extracted question sentence data in a question sentence database;
Performing the control so that the step of collecting, the step of extracting, and the step of storing are performed at a predetermined timing;
Receiving the search request of the user terminal or latte text data,
Retrieving a previous SL plurality of first question data relating to the text data for all question data stored in the question database,
Transmitting a page including a list of the retrieved first question sentence data and a selection unit to the user terminal;
Receiving a search request one of the first question data selected by said selecting means from the first question data Oite of the plurality to said user terminal,
Retrieving the previous SL all questions of the one of text data received by targeting a plurality of further related to the first question data second question data stored in the question database,
Transmitting a page including a list of the plurality of second question sentence data searched and a selection unit to the user terminal;
When the user terminal further receives a search request for one second question sentence data selected by the selection means from the plurality of second question sentence data, the one second question sentence data is As the first first question sentence data, repeatedly executing the searching step and the transmitting step;
Acquiring response data for the received question sentence data from the content server when receiving an answer request for the first question sentence data or the first second question sentence data from the user terminal; ,
Transmitting the obtained answer data to the user terminal together with the received question sentence data;
Including methods. A program for causing a server connected to a network to create a question sentence database and transmitting question sentence data stored in the question sentence database to a user terminal connected to the network,
Accumulating content data from a content server connected to the network according to URL list data set in advance;
Extracting the question sentence data that matches the extraction condition from the accumulated content data according to a preset extraction condition;
Storing the extracted question sentence data in a question sentence database;
Performing the control so that the step of collecting, the step of extracting, and the step of storing are performed at a predetermined timing;
Receiving the search request of the user terminal or latte text data,
Retrieving a previous SL plurality of first question data relating to the text data for all question data stored in the question database,
Transmitting a page including a list of the retrieved first question sentence data and a selection unit to the user terminal;
Receiving a search request one of the first question data selected by said selecting means from said plurality of first question data at the user terminal,
Retrieving the previous SL all questions of the one of text data received by targeting a plurality of further related to the first question data second question data stored in the question database,
Transmitting a page including a list of the plurality of second question sentence data searched and a selection unit to the user terminal;
When the user terminal further receives a search request for one second question sentence data selected by the selection means from the plurality of second question sentence data, the one second question sentence data is As the first first question sentence data, repeatedly executing the searching step and the transmitting step;
Acquiring response data for the received question sentence data from the content server when receiving an answer request for the first question sentence data or the first second question sentence data from the user terminal; ,
Transmitting the obtained answer data to the user terminal together with the received question sentence data;
A program for causing the server to execute.

Images