JP2016192067A - Searcher device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a searcher device capable of presenting a candidate word for performing an appropriate search.SOLUTION: Search condition acquisition means 2 acquires a search condition specified by a user. Search execution means 4 searches a piece of target information 12 stored in a storage section 10 using the acquired search condition. Search result output means 6 outputs a piece of search information 14 as a search of the result. Candidate word output means 8 extracts a feature word from the search information 14 and outputs the same as a candidate word. The candidate word output means 8 may extract a feature word from the object information 12 not from the search information 14 and output the same as a candidate word.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a search device for searching for target information, and more particularly to a search word suggestion function.

  When searching for target information, there are cases where the search condition is not sufficiently narrowed down and the number of hits is too large, or conversely, the number of hits becomes zero due to excessive narrowing down.

  When the narrowing down is not sufficient, the user who performs the search adds a new search term to the search condition and performs the narrowing down. If the number of hits is too small, the user who performs the search discards the search result based on the search term and performs a search using another search term. In any case, selecting an appropriate search term is important for obtaining favorable search results.

  Patent Document 1 discloses a device that displays candidates for a co-occurrence word of a search word included in a search formula as a new search word when the number of hits is too large.

  Patent Document 2 discloses a system for presenting a product advertisement of an appropriate category to a website. A category name is searched as a search word, a feature word is extracted from the search result, and this feature word is associated with the category name to generate a feature word database. For websites that contain many feature words in the feature word database, product advertisements in the associated category are put out.

JP2002-230037 JP2010-009307

  However, the technique disclosed in Patent Document 1 displays a co-occurrence word of a search word as a candidate, so that there is a problem that a candidate cannot be displayed when there is no co-occurrence word.

  In addition, there is a problem that appropriate narrowing down cannot be performed unless the co-occurrence word is included in the retrieved information.

  On the other hand, in the technique of Patent Document 2, although an advertisement of an appropriate category is realized by the feature word database, it is not recognized as a problem to appropriately narrow down the search.

  An object of the present invention is to provide an apparatus that solves the above problems and presents candidate words for performing an appropriate search.

  Some features that can be applied independently of the input device according to the present invention will be described below.

(1) (2) A search device according to the present invention includes a search condition acquisition means for acquiring a search condition by a user, a search execution means for searching for target information based on the acquired search condition, and a search searched by the search execution means Search result output means for outputting information, and candidate word output means for extracting feature words included in the search information or the target information and outputting them as candidate words for search.

  Therefore, an appropriate candidate word for further search can be output.

(3) In the search device according to the present invention, the candidate word output means narrows down the search information or target information based on at least the search date, search place, or user attribute, and the narrowed search And a feature word extracting means for extracting a feature word included in the information or the target information.

  Therefore, more appropriate candidate words can be output.

(4) The search device according to the present invention is characterized in that the narrowing down means narrows down the search information or the target information to those for which the user has access authority.

  Therefore, information without access authority can be excluded from the feature word extraction, and more appropriate candidate words can be output.

(5) In the search device according to the present invention, the candidate word output means includes a feature word based on the search information or information obtained by narrowing down the search information, and a feature word based on the target information or information obtained by narrowing down the target information. Both are output as candidate words.

  Accordingly, selectable candidate words can be enriched.

(6) The search device according to the present invention is characterized in that the search condition acquisition means acquires the search condition from a terminal device connected via a network.

  Therefore, a search device that can be accessed from a plurality of terminal devices can be realized.

(7) A search method according to the present invention is a search method by a computer, wherein the computer acquires a search condition by a user, the computer searches for target information according to the acquired search condition, and the computer performs a search execution means. The search information searched according to the above is output, and the computer extracts feature words included in the search information or result information searched under the search condition obtained by removing at least one search word from the search condition. Is output as a candidate word.

  Therefore, an appropriate candidate word for further search can be output.

  In the embodiment, “search condition acquisition unit” corresponds to step S2.

  In the embodiment, “search execution means” corresponds to step S3.

  “Search result output means” corresponds to step S4 in the embodiment.

  In the embodiment, “candidate word output means” corresponds to steps S7, S9, and S10.

  In the embodiment, “squeezing means” corresponds to steps S72 and S92.

  “Feature word extraction means” corresponds to steps S7, S9, S74, and S94.

  “Target information” refers to information to be searched. Data formats such as text data and image data are not limited.

  “Search information” refers to result information obtained by searching target information. Note that when the search information of this time is further searched, the search information of this time becomes the target information.

  The “program” is a concept that includes not only a program that can be directly executed by the CPU, but also a source format program, a compressed program, an encrypted program, and the like.

It is a functional block diagram of the search device by one Embodiment of this invention. This is a system configuration when the search device is configured as the server device 30. It is a hardware configuration of the server device 30. 5 is a flowchart of a search program 54. 5 is a flowchart of a search program 54. It is an example of data of user DB. It is an example of data of DB used as search object. It is an example of the search result display screen in the terminal device. It is a figure which shows the table of appearance frequency. It is an example of a display of the candidate word in the terminal device 20. It is a display example of candidate word selection in the terminal device 20. It is a functional block diagram of the search device by a 2nd embodiment. 5 is a flowchart of a search program 54. 5 is a flowchart of a search program 54. It is an example of search result data.

1. First embodiment
1.1 Overall Configuration FIG. 1 is a functional block diagram of a search device according to an embodiment of the present invention. The search condition acquisition unit 2 acquires a search condition specified by the user. The search execution means 4 searches the target information 12 recorded in the recording unit 10 using the acquired search conditions. The search result output means 6 outputs search information 14 that is a search result.

The candidate word output means 8 extracts the feature word of the search information 14 and outputs it as a candidate word. Note that the candidate word output unit 8 may extract a feature word of the target information 12 instead of the search information 14 and output it as a candidate word.

1.2 System Configuration FIG. 2 shows a system configuration when the search device is configured as the server device 30. A terminal device 20, 20,..., And a database device 32 are connected to the server device 30 as a search device via the Internet / LAN. The database device 32 collects and accumulates information from other server devices connected via a LAN, a web server on the Internet, and the like. In addition, a user DB describing user attributes (name, file browsing authority, etc.) is also recorded.

  Therefore, the user of the terminal device 20 can search for a file recorded in the database device 32 via the server device 30.

  FIG. 3 shows a hardware configuration of the server device 30. A memory 42, a hard disk 46, a DVD-ROM drive 48, and a communication circuit 50 are connected to the CPU 40.

  The communication circuit 50 is a circuit for connecting to a LAN or the Internet. An operating system 52 and a search program 54 are recorded on the hard disk 46. The search program 54 performs its function in cooperation with the operating system 52. These programs are those recorded on the DVD-ROM 56 and installed on the hard disk 46 via the DVD-ROM drive 48.

The hardware configuration of the terminal device 20 is the same as that in FIG. However, in the terminal device 20, the browsing program is recorded in the hard disk 46 instead of the search program 54. A keyboard, mouse, display, etc. are also connected.

1.3 Search Process FIGS. 4 and 5 show flowcharts of the search program 54. First, the CPU 40 of the server device 30 (hereinafter may be abbreviated as the server device 30) receives a login request from the terminal device 20 and performs a login process (step S1). In the login process, authentication is performed based on the user ID and password sent from the terminal device 20. FIG. 6 shows an example of the user DB recorded in the DB device 32. The server device 30 performs authentication with reference to this user DB.

  Next, the server device 30 receives and acquires search conditions from the user who operates the terminal device 20 (step S2). It is assumed that the search condition “product * manual” is received as the current search condition. Note that “*” means AND, and the terms “product” and “manual” are search conditions for searching for files included in the text. Here, it is assumed that the search has already been performed for “product”.

  Subsequently, the server device 30 searches the file recorded in the DB device 32 according to the above search conditions (step S3). FIG. 7 shows files recorded in the DB device 32. These files are regularly acquired by the DB device 32 from another server device (for example, personnel database) via the in-house LAN. The file includes a page in HTML format.

  As shown in FIG. 7, a content ID, a title, a file name, a body text, an update date and time are recorded. In the referenceable time, a time that can be referred to by each user is recorded. For example, the “labor union page” of the content (ID = 4) can be referred to between 18:00:00 and 23:00, which is not a regular workday.

  The server device 30 transmits the search result to the terminal device 20 (step S4). The terminal device 20 displays this on the display. FIG. 8 shows a search result screen displayed on the terminal device 20.

  A content ID, a title, and a part of the text are shown. The user can request the entire text to the server device 30 by operating the terminal device 20 and clicking on a part (linked) of the text.

  The user who operates the terminal device 20 selects whether to further narrow down the search result or to narrow down with another search word instead of the current search word. This selection is performed by clicking the narrowing method button 60 or 62 in FIG. If the button 60 is clicked, the search result can be further narrowed down. This is effective when the number of search results is large. If the button 62 is clicked, it is possible to narrow down by another search word instead of the current search word (“manual”). This is effective when the number of search results is too small.

  Here, the description will be made assuming that the button 60 (AND) is clicked. The terminal device 20 transmits to the server device 30 that the button 60 has been clicked. The server device 30 acquires a narrowing method from the terminal device 20 (step S5).

  The server device 30 determines whether or not the narrowing-down method transmitted from the terminal device 20 is AND (step S6). Here, since it is AND, step S7 is executed. In step S7, the server device 30 extracts feature words included in each file searched in step S3 (step S7).

  In this embodiment, a feature word is obtained by performing morphological analysis on the body of each file. Further, the appearance frequency is calculated for each feature word. The obtained feature words and the number of appearances are shown as an example in FIG. As shown in FIG. 9, the server device 30 also calculates and records the degree of importance (ratio% of the number of appearances) when the largest number of appearances is set to 100. The server device 30 extracts the top three (predetermined number) feature words of appearance frequency. Therefore, “sales promotion material”, “product price”, and “inquiry” are extracted here.

  Next, the server device 30 determines whether or not the narrowing-down method transmitted from the terminal device 20 is OR (step S8). Here, since it is AND, step S10 is performed. In step S <b> 10, the server device 30 transmits the extracted feature word as a candidate word to the terminal device 20.

  In response to this, the terminal device 20 displays the candidate words on the display. FIG. 10 shows a display example on the terminal device 20. Candidate words are shown in the candidate word display column 64 together with their importance.

  When the user operates the terminal device 20 to select a displayed candidate word, the terminal device 20 transmits the selected candidate word to the server device 30. For example, when the user selects “sales promotion material”, this is transmitted to the server device 30.

  In response to this, the server device 30 acquires a new search expression (product * manual * promotional material) (step S2). The server device 30 executes step S3 and subsequent steps again. FIG. 11 shows a search result screen. In this way, an appropriate refinement search can be performed by presenting an appropriate candidate word.

  In the above embodiment, the case where the button 60 (AND) is selected has been described. The processing when the button 62 (OR) is selected on the screen of FIG. 8 is as follows. The button 62 (OR) is used when the current search result is not preferable or when the number of search results is too small.

  In step S6, the server device 30 determines whether the narrowing-down method is AND. Here, since it is OR, it progresses to step S8. In step S8, the server device 30 determines whether or not the narrowing-down method is OR. Here, since it is OR, it progresses to step S9.

  In step S <b> 9, the server device 30 acquires a feature word for the file searched by the previous search. In the case shown in FIG. 8, since “product” was searched as a search word in the previous search, a feature word is obtained by the same method as in step S7 for the file that is the search result, and appears. The top three are extracted according to the number of times (step S9). In view of such processing, it is preferable that the search results up to the previous time be stored and used without being discarded.

  Subsequently, the server device 30 transmits the extracted feature words to the terminal device 20 as candidate words. The terminal device 20 displays this candidate word in the candidate word display field 64. When a candidate word is selected by the user, the server device 30 generates a new search expression based on the selected candidate word instead of the search word added in the current search, and executes the search (step S2 and subsequent steps). ).

In this way, the re-search can be performed so as to appropriately narrow down the previous search result. If there is no previous search result, the process of step S9 is not performed.

1.4 Other
(1) In the above embodiment, the user who operates the terminal device 20 selects the narrowing-down method (AND or OR). However, it may be set in advance by default. In this case, the user does not need to make a selection if the default is satisfied. Select only if different from the default.

  Further, the server device 30 may automatically select based on the number of search hits. For example, if the number of search results is equal to or greater than a predetermined number (for example, 1000 or more), the possibility of further narrowing down is high, so AND is selected. Further, when the number of search results is less than a predetermined number (for example, 10), it is determined that it is necessary to narrow down with other search terms, and OR is selected. If none of these apply, the user makes a selection.

(2) In the above embodiment, AND and OR are adopted as the narrowing down method. However, NOR, NAND, XOR, etc. may be adopted.

(3) In the above embodiment, feature words are extracted from the previous search result file in step S9. For example, if the current search expression is (A * B * C) and the previous search expression is (A), the search expression (A) (that is, a file including the search word (A) is searched) Extract feature words from search results.

  However, feature words may be extracted from a search result file based on a search expression obtained by reducing one search word. For example, in the above case, feature words may be extracted from the search result of the search expression (A * B).

(4) In the above embodiment, feature words are extracted from the search result (search information) in step S7. However, feature words may be extracted from the previous search result (target information).

(5) In the above embodiment, feature words are extracted from the previous search result (target information) in step S9. However, feature words may be extracted from search results (search information).

(6) In the above embodiment, feature words are extracted from search information or target information and output as candidate words. However, feature words may be extracted from both search information and target information and output as candidate words.

(7) In the above embodiment, feature words are extracted based on the appearance frequency of words in search information or target information. However, the feature word may be extracted based on whether it is designated as a title by a tag or the like, whether it is designated as bold, or the like. In addition, feature words may be extracted by combining these methods.

  Also, instead of summing up the number of appearances in each file, the ratio of the number of appearance ranks of the word (appearance ratio) in the total number of words is calculated for each file, and the appearance ratio for each file is summed up ( The feature words may be extracted based on the total appearance ratio).

(8) In the above embodiment, the search device is configured as the server device 30 and receives input from the user via the network. However, the search device itself may directly receive input from the user and display the search result.

(9) Each of the above modifications can be applied to the second embodiment and other embodiments as long as the essence is not violated.

2. Second embodiment
2.1 Overall Configuration FIG. 12 shows a functional block diagram of a search device according to the second embodiment of the present invention. The search condition acquisition unit 2 acquires a search condition specified by the user. The search execution means 4 searches the target information 12 recorded in the recording unit 10 using the acquired search conditions. The search result output means 6 outputs search information 14 that is a search result.

The narrowing-down means 80 of the candidate word means 8 narrows down the search information 14 based on a predetermined user attribute or the like. For example, the search information is narrowed down to only search information 14 for which the user has browsing authority. The feature word extraction unit 82 extracts feature words of the narrowed search information 14. The output unit 84 outputs the extracted feature word as a candidate word. The narrowing means 80 may narrow down the target information 12 instead of the search information 14.

2.2 System Configuration The system configuration and hardware configuration are the same as those in the first embodiment.

2.3 Search Process FIGS. 13 and 14 show flowcharts of the search program 54. Steps S1 to S5 shown in FIG. 13 are the same processes as in the first embodiment.

  In step S6, the server device 30 determines whether the narrowing-down method designated by the user is AND. If the narrowing-down method is AND, the search result is narrowed down by the access authority of the user (step S72).

  For example, FIG. 15 shows the searched files for the search results shown in FIG. The server device 30 narrows down only those files for which the user has access authority from among the retrieved files. For example, if the user operating the terminal device 20 is user01, only the contents in which user01 is described in the access authority column are narrowed down. In the case of FIG. 15, the content ID = 1, 2, 3, 4,... Is narrowed down.

  Next, the server device 30 extracts feature words from the narrowed search result file (step S74). This feature word extraction processing is the same as in the first embodiment.

  In step S8, the server device 30 determines whether the narrowing-down method designated by the user is OR. If the narrowing-down method is OR, the previous search result is narrowed down by the access authority of the user (step S92). The narrowing-down method is the same as in step S72.

  Next, the server device 30 extracts feature words from the narrowed search result file (step S94). This feature word extraction processing is the same as in the first embodiment.

  Thereafter, the server device 30 transmits the extracted feature words as candidate words to the terminal device 20 (step S10). The terminal device 20 displays this.

As described above, according to this embodiment, since feature words are extracted from search results for which the user has access authority, it is possible to perform a narrowing-down that matches the user.

2.4 Other
(1) In the above embodiment, narrowing down is performed by access authority. However, narrowing down may be performed using user attributes other than access authority. For example, a flag for content for women and content for men may be recorded on the file side, and narrowed down according to the gender of the user.

  Further, as shown in FIG. 15, the referable time may be recorded in the DB, and narrowing down may be performed depending on whether or not the time when the user searches is within the time. Record the location that can be referenced (whether it can be referred only from the terminal device in the company or from the terminal device outside the company) in the file, and narrow down based on the location of the terminal device that the user searched for You may do it.

  In addition, narrowing down may be performed according to file size, file type, file update date and time, and the like.

(2) In the above embodiment, narrowing down is performed for extracting feature words. However, the same narrowing may be performed in the search result output. For example, a file for which the user does not have access authority may not be included in the search result and may not be displayed.

(3) In the above embodiment, search information (current search result) or target information (previous search result) is narrowed down by user attribute, etc., feature words are extracted from the narrowed down information, and output as candidate words. Yes.

  However, search information is narrowed down by user attribute, etc., and feature words are extracted from the narrowed down information, target information is narrowed down by user attribute, etc., feature words are extracted from the narrowed down information, and both are output as candidate words. May be.

  In addition, both feature words for search information and feature words for information obtained by narrowing down target information based on user attributes may be output as candidate words. Furthermore, both feature words for information obtained by narrowing search information by user attributes and the like and feature words for target information may be output as candidate words.

(4) In the above embodiment, the server device 30 automatically narrows down the user attributes. However, the user may instruct which attribute, date, etc., to narrow down, and the narrowing may be performed in response to this.

(5) Each of the above-described modifications can be applied to the first embodiment and other embodiments as long as the essence is not violated.

Claims (7)

Search condition acquisition means for acquiring a search condition by a user;
A search execution means for searching for target information according to the acquired search condition;
Search result output means for outputting search information searched by the search execution means;
Candidate word output means for extracting feature words included in the search information or the target information and outputting as candidate words for search;
Search device equipped with. A search program for realizing a search device by a computer, comprising:
Search condition acquisition means for acquiring a search condition by a user;
A search execution means for searching for target information according to the acquired search condition;
Search result output means for outputting search information searched by the search execution means;
A search program for extracting a feature word included in the search information or the target information and functioning as candidate word output means for outputting as a candidate word for search. In the apparatus of claim 1 or the program of claim 2,
The candidate word output means includes:
Narrowing means for narrowing down the search information or the target information based on at least a search date, a search location, or a user attribute;
Feature word extraction means for extracting feature words included in the refined search information or target information;
A device or program comprising: In the apparatus or program of Claim 3,
The narrowing-down means narrows down the search information or the target information to those for which the user has access authority. In the apparatus or program in any one of Claims 1-4,
The candidate word output means outputs both the feature word based on the search information or information narrowed down the search information and the feature word based on the target information or information narrowed down the target information as candidate words. A device or program characterized by the above. In the apparatus or program in any one of Claims 1-5,
The search condition acquisition means acquires a search condition from a terminal device connected via a network. A computer search method,
The computer gets the search conditions by the user,
The computer searches the target information according to the acquired search conditions,
The computer outputs the search information searched by the search execution means,
The computer extracts feature words included in the search information or search result information obtained by removing at least one search word from the search conditions, and outputs the extracted feature words as candidate words. How to search.

Images