JP6204266B2 - Information retrieval device - Google Patents

Description

  The present invention relates to an information search apparatus that supports processing for searching for information desired by a user from information on a network.

  In recent years, with the spread of the Internet, search engines for searching for desired information are widely used.

  For information retrieval by a search engine, for example, there are support methods based on processes such as the following (1) and (2).

  (1) Processing for presenting a recommended query reflecting the user's interest to the user in a keyword search in which the user inputs a search term called a “query” into the search window to search (Patent Document 1).

  (2) Create a hierarchical site by classifying information for a certain category, and display the related classification and hierarchy link destinations on the content being browsed based on the site, and click operation To transition to the linked page.

  By using such processing, the user can browse desired information from a huge amount of information.

JP2012-59182A

  Normally, when an information search is started, the intention of the user's search is not clear, and it is often embodied in the process of performing the search process as described above.

  For example, when a user searches for information related to “food”, the user first searches for information related to “pasta” only by thinking that “I want to eat pasta”, and compares the output information. It is conceivable that the search is advanced in order to acquire information on "".

  In the course of the search, the information intended by the user is embodied as “food” → “pasta” → “cream pasta”. The level of abstraction related to the category is getting lower. At that time, in order to enable the user to more specifically specify the intention of the search, it is desirable to present information of the same abstraction level as much as possible without duplication.

  However, although various options are presented in the search processes (1) and (2) as described above, the information presented in the course of the search is not presented because the information is not presented in consideration of the degree of abstraction. The range of information provided due to duplication may be narrowed. In that case, since the user will compare the intention within the narrow range of information and embody the intention of the search, there is a problem that it may not be possible to obtain high quality information that matches the user's intention. .

  In addition, even when the above-described search process is used, if the search process is performed again with a keyword indicating the intention embodied by advancing the search, there is a possibility that high-quality information specialized for the keyword can be acquired. However, there is a problem in that it takes time because it is necessary to re-enter the keyword.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide an information search device that efficiently and accurately searches information corresponding to a search target intended by a user from a huge amount of information. And

In order to solve the above problems, an information search apparatus of the present invention includes a content DB that holds a plurality of content information to be provided to a user, and a plurality of words each having a plurality of words for each set category. The category hierarchy information DB that holds the category hierarchy information that is hierarchically configured so that the level of abstraction of the semantic attribute of the label decreases as the level of the node increases from the higher level to the lower level, and the keyword information input by the user's operation Alternatively, based on the input information acquisition unit that acquires the address information of the content information as user input information, and the user input information acquired by the input information acquisition unit, the search target from among the information held in the category hierarchy information DB identify the category hierarchy information, a no from a particular category hierarchy information specified based on the acquired user input information Is associated with the user input information, the parent node of the associated node is also associated with the user input information, and there is a node having a plurality of associated child nodes among the associated nodes. If the node with the highest level of abstraction is specified as the search intention node that is the node of the label that the user intends to search for, and there is no node having a plurality of associated child nodes A search intention specifying unit that specifies a node having the lowest abstraction level among the associated nodes as a search intention node, and a search key specified based on the search intention node specified by the search intention specifying unit. A search processing unit that acquires search results searched in the content DB, and a display unit that displays the search results acquired by the search processing unit. It is characterized in.

  In addition, the search processing unit of the information search device searches the content DB using the labels of all the child nodes of the search intention node as keys, and acquires content information including each key as a search result. Good.

  The display unit of the information search device displays the category hierarchy information to be searched, and the search processing unit stores the content in the content DB based on a node specified from the category hierarchy information displayed on the display unit. The search result may be acquired as the search result.

  According to the information search device of the present invention, it is possible to efficiently obtain information corresponding to a search target intended by a user when performing a search process from an enormous amount of information.

It is a block diagram which shows the structure of the information search device by one Embodiment of this invention. It is explanatory drawing which shows an example of the category hierarchy information hold | maintained at category hierarchy information DB of the information search device by one Embodiment of this invention. It is a flowchart which shows the search operation | movement in the information search device by one Embodiment of this invention. It is explanatory drawing which shows the node by which matching was performed in the category hierarchy information of the category hierarchy information DB of the information search device by one Embodiment of this invention. It is a flowchart which shows the specific operation | movement of the search intent node in the information search device by one Embodiment of this invention. It is an example of the screen block diagram which shows the search result displayed by the information search device by one Embodiment of this invention. It is explanatory drawing which shows the node by which matching was performed in the category hierarchy information of the category hierarchy information DB of the information search device by one Embodiment of this invention. It is an example of the screen block diagram which shows the search result displayed by the information search device by one Embodiment of this invention. It is another example of the screen block diagram which shows the search result displayed by the information search device by one Embodiment of this invention.

<Configuration of Information Retrieval Device According to One Embodiment>
A configuration of an information search apparatus according to an embodiment of the present invention will be described with reference to FIG.

  The information search apparatus 1 according to the present embodiment includes a content DB 11, a category hierarchy information DB 12, an input information acquisition unit 13, a search intention specifying unit 14, a search processing unit 15, and a display unit 16.

  The content DB 11 holds various Web page information and its address information to be provided to the user as content information.

  For each set category, the category hierarchy information DB 12 is configured in a hierarchical manner so that a plurality of nodes each having a plurality of words as labels are hierarchically arranged so that the level of abstraction of the semantic attribute of the label decreases from the top to the bottom. Holds category hierarchy information.

  The input information acquisition unit 13 acquires keyword information or Web page address information input by a user operation as user input information.

  Based on the user input information acquired by the input information acquisition unit 13, the search intention specifying unit 14 specifies the category hierarchy information to be searched from those stored in the category hierarchy information DB 12, and from the specified category hierarchy information, A search intention node which is a node of a label intended by the user as a search target is specified using the abstraction level of the semantic attribute of the label.

  The search processing unit 15 searches the content information in the content DB 11 based on the user input information input from the input information acquisition unit 13, and all the child nodes of the search intention node determined by the search intention specifying unit 14. The content information in the content DB 11 is searched using the label as a search key, and the link information of the corresponding content information is acquired as the search result.

  The display unit 16 is composed of a monitor, and displays link information of content information acquired as a search result by the search processing unit 15.

<Operation of Information Retrieval Device According to One Embodiment>
Next, the operation of the information search device 1 according to the present embodiment will be described.

  In the present embodiment, the content DB 11 includes a plurality of pieces of content information including words such as “animal”, “cat”, “dog”, “fish”, “Abinician”, “Persia”, “Spitz”, and “Chihuahua”. Is retained.

  The category hierarchy information DB 12 holds category hierarchy information for each category. For example, the category hierarchy information of the category <animal> as shown in FIG. 2 is the “animal” node 20 labeled “animal”. “Cat” node 21 labeled “cat” as a child node, “dog” node 22 labeled “dog”, “bird” node 23 labeled “bird”, “fish” labeled “ There are a plurality of nodes such as a “fish” node 24, and a plurality of nodes such as an “Abinician” node 25 labeled “Abinician” as a child node of the “cat” node 21 and a “Persian” node 26 labeled “Persia”. There are two nodes, "Spitz" node 27 labeled "Spitz" as a child node of "Dog" node 22, "Chihuahua" node 28 labeled "Chihuahua", etc. As more nodes are present, it is hierarchically configured.

  In the category hierarchy information of the category <animal> in FIG. 2, the “animal” node 20 belonging to the uppermost layer is defined as a node in the first hierarchy, and then the “cat” node 21 and “dog” belonging to the upper hierarchy. The node 22, the “bird” node 23, the “fish” node 24, etc. are defined as nodes in the second hierarchy, and the “Abinician” node 25, “Persia” node 26, “Spitz” node 27, “ The “chihuahua” node 28 and the like are defined as nodes in the third hierarchy. Each node is associated with an abstraction level of the semantic attribute of the label in the category <animal>, and this abstraction level is indicated by a lower value as the level of the node becomes lower. For example, the “animal” node 20 in the first hierarchy is associated with an abstraction level “3”, and the “cat” node 21, “dog” node 22, “bird” node 23, and “fish” node 24 in the second hierarchy. Is associated with an abstraction level “2”, and an “Abinician” node 25, a “Persia” node 26, a “Spitz” node 27, and a “Chihuahua” node 28 in the third hierarchy are associated with an abstraction level “1”. ing. That is, as the first hierarchy → second hierarchy → third hierarchy and the hierarchy becomes deeper from the upper level to the lower level, the level of abstraction of the semantic attribute of the label for each node decreases.

  FIG. 3 shows the operation of the information search apparatus 1 when the user wants to search for “animal” (search intention) and the keyword information “animal” is input as user input information in a state where these pieces of information are held. This will be described with reference to the flowchart of FIG.

  First, in the input information acquisition unit 13, keyword information “animal” acquired as user input information by a user operation is acquired (S <b> 1).

  When the keyword information “animal” is acquired, the search processing unit 15 searches the content DB 11 based on the keyword information “animal” and displays the corresponding content information in the same manner as a normal search engine. Information is acquired as a first search result (S2).

  Next, the search intention specifying unit 14 determines whether or not a plurality of words are included in the text corresponding to the acquired user input information (S3). Here, since the acquired user input information is the keyword information “animal” composed of a single word and does not include a plurality of words (“NO” in S3), the process proceeds to step S4.

  In step S4, the search intention specifying unit 14 searches the category hierarchy information DB 12 for a node having the acquired keyword information “animal” as a label, and associates the node specified by the search with the keyword information “animal”. Hereinafter, the process of associating a node in the category hierarchy information DB 12 with a word related to user input information is referred to as “association”.

  As a result of the search, as indicated by a thick line in FIG. 4, the node 20 having the label “animal” in the category “animal” is associated, and the category “animal” is identified as the search target category ( S4).

  Next, the search intention specifying unit 14 executes a process of specifying a search intention node that is a label intended as a search target in the category hierarchy information of the category <animal> specified as the search target (S5). ). Processing for specifying the search intention node will be described with reference to the flowchart of FIG.

  First, if the parent node of the node associated in step S4 is not yet associated, it is newly associated (S11). Here, since the node 20 labeled “animal” is a root node having no parent node, there is no corresponding node.

  Next, it is determined whether there is a node having a plurality of associated child nodes among the associated nodes (S12). Here, the associated node is only the node 20 labeled “animal”, and the node 20 labeled “animal” does not have a plurality of associated child nodes (“NO” in S12). Therefore, the process proceeds to step S13, and a node 20 having a label “animal” as a search intention node is specified as a node existing in a hierarchy (lower hierarchy) having the lowest level of abstraction among the associated nodes.

  When the search intention node is specified, the process returns to the flowchart of FIG. 3, and the search processing unit 15 performs a search using the labels of all the child nodes of the specified search intention node as search keys (S6). Here, the contents DB 11 uses the labels “cat”, “dog”, “bird”, and “fish” of the child nodes 21, 22, 23, and 24 of the node 20 with the label “animal” as the search intention node as keys. The content information including each search key is acquired as a search result.

  Among the content information obtained as a search result, link information for displaying the highest content information for each label is acquired as the second search result, and is displayed on the display unit 16 together with the first search result acquired in step S2. It is displayed (S7).

  An example of a screen configuration diagram when the first search result and the second search result are displayed on the display unit 16 is shown in FIG. In the screen configuration diagram of FIG. 6, the display unit 16 has a first search result display area 161 for displaying the first search result on the left side, and a second search result display area 162 for displaying the second search result on the right side. doing.

  In the second search result display area 162, link information for displaying the content information acquired by the processes of steps S4 to S7 is displayed. For example, the content information “cat summary site”, which is the top search result when the search is performed using “cat” as a key, which is the label of the child node 21 of the word “animal” that the user intends to search is displayed. Link information 1621 for displaying content information “dog illustration book” that is the highest-level search result when “dog” that is the label of child node 22 is searched as a key. Link information 1623 for displaying content information “bird encyclopedia”, which is the highest-level search result when “bird” as a label is searched for as a key, and “fish” as a label for child node 24 as a key The link information 1624 for displaying the content information “fish encyclopedia” which is the highest-level search result at the time of being displayed is displayed.

  As described above, the keyword information input by the user is based on the first search result similar to the search result by the normal search engine and the word having the same abstract attribute with respect to the word intended by the user as the search target. By displaying the acquired second search result, it is possible to assist the user in selecting information with high accuracy in accordance with the abstraction level of the semantic attribute of the word when the user intends to specify the search intention.

  Next, the operation when the user clicks the link information 1621 for displaying “Cat Summary Site” in order to know “Cat” from the search results displayed as shown in FIG. 6 will be described. .

  When the link information 1621 is clicked by the user, the input information acquisition unit 13 acquires the address information of the content information “cat summary site” associated with the link information 1621 as user input information (S1).

  When the address information of the content information “cat summary site” is acquired, the search processing unit 15 searches the content DB 11 based on the address information, and the corresponding content information “cat summary site” is the first search result. Obtained (S2).

  Next, the search intention specifying unit 14 determines whether or not a plurality of words are included in the text information in the acquired content information “Cat Summary Site” (S3). Here, it is determined that a plurality of words are included in the text information in the acquired content information “Cat Summary Site” (“NO” in S3), and the text in the site is divided into words (S9).

  At this time, the terminology may be deleted from the text information based on the terminology dictionary created in advance during system configuration. By deleting the non-terms, it is possible to shorten the calculation time for specifying the search target category in the next step S4.

  It is assumed that the words “Abinician” and “Persian” are acquired by dividing the text information in the content information “Cat Summary Site” in step S9.

  Next, for all the words obtained by the division, nodes having the same word / phrase as a label are searched from the category hierarchy information DB 12, and nodes 25 and “Persia” having the label “Abinician” in the category <animal> are selected. A node 26 as a label is associated. Then, the category having the largest number of associated nodes is specified as the search target category, and here, the category <animal> is specified as the search target category (S4).

  Next, the search intention specifying unit 14 makes a node 21 labeled “Cat” that is a parent node of the node 25 labeled “Abinician” and the node 26 labeled “Persia” already associated with each other. A node 20 having the label “animal” as a parent node is also associated (S11). The nodes associated so far are indicated by bold lines in FIG.

  Next, it is determined whether there is a node having a plurality of associated child nodes among the associated nodes (S12). Here, since the node 21 having “cat” as a label corresponds (“YES” in S12), the process proceeds to step S14, and the node having a plurality of associated child nodes has the highest level of abstraction. An existing node, that is, a node 21 labeled “cat” is identified as a search intention node (S14).

  Next, a search in the content DB 11 is performed using the labels “Abinician” and “Persia” of all the child nodes 25 and 26 of the node 21 whose label is the specified search intention node “cat”, respectively. The content information including each search key is acquired as a search result (S6).

  Among the content information obtained as a search result, link information for displaying the highest content information for each label is acquired as the second search result, and is displayed on the display unit 16 together with the first search result acquired in step S2. It is displayed (S7).

  An example of a screen configuration diagram when the first search result and the second search result are displayed on the display unit 16 is shown in FIG. 8 has a first search result display area 161 for displaying the first search result on the left side of the display unit 16, and a second search result display area 162 for displaying the second search result on the right side. doing.

  In the second search result display area 162, content information “”, which is the highest search result when the search key is “Abinician”, which is the label of the child node of the word “cat” that the user intends to search, is displayed. The link information 1625 for displaying the “Abinisian specialized site” and the link information 1626 for displaying the content information “Persian Encyclopedia” which is the highest search result when “Persia” is used as a search key are displayed. Yes.

  Further, when the search result is displayed on the display unit 16, as shown in FIG. 9, display information 1627 schematically showing the category hierarchy information to be searched is also displayed and designated by the user from the displayed information. The search result may be obtained by searching the content DB 11 using the node label as a search key, and displayed. By displaying in this way, the user can specify a desired node to be searched while visually recognizing the abstraction level of each node in the category hierarchy information, and the search efficiency can be further improved. it can.

  According to the present embodiment described above, when a search process is performed from an enormous amount of information, it is possible to efficiently obtain information corresponding to the search target intended by the user.

  In the above-described embodiment, the case where the first search result and the second search result are displayed on the same screen has been described. However, the present invention is not limited to this, and the mode in which the first search result is displayed on the entire screen and the second mode. You may comprise so that the mode which displays a search result on a full screen can be switched.

DESCRIPTION OF SYMBOLS 1 ... Information retrieval apparatus 11 ... Content DB
12 ... Category hierarchy information DB
DESCRIPTION OF SYMBOLS 13 ... Input information acquisition part 14 ... Search intention specific part 15 ... Search process part 16 ... Display part 20-28 ... Node 161 ... 1st search result display area 162 ... 2nd search result display area 1621-1626 ... Link information 1627 ... Display information schematically showing category hierarchy information

Claims (3)

A content DB that holds a plurality of content information to be provided to the user;
For each set category, a plurality of nodes each having a plurality of words as labels are held in a hierarchically structured category hierarchy information so that the abstraction level of the semantic attribute of the label becomes lower from the upper level to the lower level. Category hierarchy information DB;
An input information acquisition unit that acquires, as user input information, address information of keyword information or content information input by a user operation;
Based on the user input information obtained by the input information acquisition unit, specifies the search target category hierarchy information from said retained in the category hierarchy information DB, from the specified category hierarchy information, the obtained user input The node identified based on the information is associated with the user input information, the parent node of the associated node is also associated with the user input information, and a plurality of associated child nodes are associated among the associated nodes. If there is a node that has, the node with the highest level of abstraction is identified as the search intention node that is the label node that the user intends to search and has a plurality of associated child nodes If no node exists, the node with the lowest level of abstraction among the associated nodes is specified as the search intention node. And search intent specifying unit,
A search processing unit for acquiring a search result obtained by searching the content DB with a search key specified based on a search intention node specified by the search intention specifying unit;
An information search apparatus comprising: a display unit that displays a search result acquired by the search processing unit. The search processing unit searches the content DB using labels of all child nodes of the search intention node as keys, and acquires content information including each key as a search result. The information search device according to claim 1. The display unit displays the category hierarchy information of the search target,
The search processing unit acquires, as a search result, a result of searching the content DB based on a node designated from the category hierarchy information displayed on the display unit. The information search device according to claim 1 or 2.

Images