JP2013206434A - Time condition presentation method and device and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suggest a time consideration in consideration of an appropriate time suggestion width different for each keyword (topic).SOLUTION: In the present invention, when determining a time condition on the basis of time information of a document retrieved on the basis of a retrieval condition, a time width of the document is obtained from the time information of the document, the most frequent time width of the document is turned to the width of the time condition to be suggested, and a score to be higher for the document for which the time width of the document is closer to the width of the time condition to be suggested is imparted to each time unit in a time range of the document. The scores are summed up for the respective time units, and the time range of the highest score in the time range having the width of the time condition to be suggested is extracted as the time condition.

Description

  The present invention relates to a time condition presentation method, apparatus, and program, and more particularly, to a search condition suggestion technique for assisting input of a time condition in a search range.

  First, a conventional search technique that handles time as a range will be described.

  There has been considered a date and time designation search technique for searching for documents using time as a search condition. Usually, in addition to specifying the time condition, the user also performs a search using a keyword, and the search system presents a document that satisfies both the specified time condition and the specified keyword. At this time, unlike the search based on keyword matching, the search condition time is treated as a condition having a range. For example, when a time outside the search range is specified, such as “August to September 20010”, the date and time expression included in the document, the date when the document was collected, etc. are used. Documents included in “month” can be searched (for example, see Non-Patent Document 1).

  Next, a technique for presenting time conditions in information retrieval will be described.

  In a general Web information retrieval system, when information that a user wants to search is input as a keyword, the system presents a document closely related to the input keyword to the user. In such a system, when it is desired to narrow down search results using time conditions such as “2010 information”, it has been realized by inputting the keyword “2010” into the search conditions. There is a method in which the system presents the time condition to the user in order to assist the user who wants to narrow down by the time condition but does not know what time is specified and the target information can be reached. In this paper, this is called the time condition suggest function.

  In the case of suggesting time conditions, those skilled in the art set candidate time conditions as shown in FIG. 1, perform scoring on each time condition in some way, rank them, and present them. There will be a naive way. If it is simply performed, each unit time handled by the system is set as a time condition candidate (for example, in a system that handles one day as a unit time, the time condition candidates are “August 11, 2011”, “August 12, 2011 The frequency of the date and time expression that appears in the document group to be searched can be given as the score of the candidate time condition corresponding to the date and time expression.

Nobuaki Makishima, Katsuto Bessho, Yoshimasa Koike, Ryoji Kataoka, "Specified date and time search considering the valid range of written date and time", WebDB Forum2010, 2010.

  However, as a time condition for suggesting, a seasonal time when a keyword has become a topic may be presented. The range of seasonal time that has become a hot topic is not necessarily constant, and varies from keyword to keyword. If the duration of the suggestion is fixed in advance as in the prior art, it is not possible to cope with a different duration for each keyword.

  The present invention has been made in view of the above points, and provides a time condition presentation method, apparatus, and program capable of suggesting a time condition that takes into consideration an appropriate time suggestion width that differs for each keyword (topic) The purpose is to do.

In order to solve the above problem, the present invention is a time condition presenting method for presenting a time condition in a range search,
Based on the input search condition, the search means refers to a document storage means that holds a keyword for each document and time information obtained by converting the time expression included in the document into the minimum unit handled by the system. A search step for retrieving time information;
A time condition determining step for determining a time condition based on the time information of the document;
A time condition output means for outputting the time condition;
And
In the time condition determining step,
Determining a document time width from the time information of the document, and a suggestion time width determination step as a time condition width for suggesting a time width of the most frequent document;
A document scoring step of assigning to each time unit in the time range of the document a score that increases as the time width of the document approaches the width of the time condition to suggest;
A time condition extracting step of summarizing scores for each time unit, and extracting a time range having the highest score as a time condition among time ranges having a range of time conditions to be suggested,
It is characterized by including.

Further, the present invention provides the suggestion time width determination step,
The difference between the minimum time and the maximum time of the time information is set as the time width of the document.

  In addition to the method for obtaining the time width of the document, the time width having the highest frequency among the time information can be set as the time width of the document.

  As described above, in the present invention, in the search that treats time as a range, instead of fixing the range of time conditions to suggest as in the conventional method, the time width corresponding to the keyword is obtained, and the time width and each By making the time condition range variable by using the degree of matching with the document, it is possible to suggest a time condition corresponding to the width of the seasonal period that has become a different topic for each search condition.

It is an example of the suggestion candidate by the conventional method. It is a block diagram of the time condition presentation apparatus in one embodiment of this invention. It is an example of the document database in one embodiment of this invention. It is a flowchart of the time condition presentation apparatus in one embodiment of this invention. It is a flowchart of the detailed process of the time condition suggestion part in one embodiment of this invention. It is an example which calculates | requires the time width of the document in one embodiment of this invention. It is an example of the time range to suggest in one embodiment of this invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 2 shows the configuration of the time condition presentation device in one embodiment of the present invention.

  The time condition presentation device 100 shown in FIG. 1 includes a query input unit 110, a search unit 120, a time condition suggestion unit 130, a time condition suggestion output unit 140, and a document DB 150.

  The document DB 150 is a storage medium such as a hard disk device that holds information for narrowing down by a search condition and time information extracted from a document group to be searched. Here, the time information refers to a time expression (such as August 2011) included in a document converted to a Julian date or the like that is aligned with the minimum unit handled by the system. The time information has a range from the minimum time to the maximum time, and the time is handled in the minimum unit handled by the system. For example, if the minimum unit of time handled by the system is one day, the time information “August 2011” is handled as August 1, 2011 to August 31, 2011. FIG. 3 shows an example of a document database. In the example shown in the figure, keywords and time information extracted from one document are stored in one record, and the time information is in the Julian date range with one day as the minimum unit. There may be a plurality of pieces of time information in one record. Note that the document database 150 including time information can be realized by a conventional technique.

  The operation in the above configuration will be described below.

  FIG. 4 is a flowchart of the time condition presenting apparatus in one embodiment of the present invention.

  Step 110) The search condition input unit 110 receives a search condition (query) from the user as an input and passes it to the search unit 120.

  Step 120) The search unit 120 acquires a record that matches the search condition received from the search condition input unit 110 from the document DB 150. Among the acquired records, the time information column is passed to the suggestion time condition suggestion unit 130. The processing of the search unit 120 can be realized by existing technology.

  Step 130) The time condition suggestion unit 130 receives the time condition from the search unit 120, determines the time condition to be suggested, and passes it to the time condition suggestion output unit 140. Detailed processing of the time condition suggestion unit 130 is shown in FIG.

  Hereinafter, the process of the time condition suggestion unit 130 will be described by dividing it into three processes of determination of a suggestion time width, document scoring, and time condition extraction. Note that the calculation results of the following processes are stored in a memory (not shown) in the time condition suggestion unit 130.

1) Suggestion time width determination process The time information is received from the search unit 120 (step 131), and the width of the time condition for suggestion is determined from the received time information. The width of the time handled by each document is obtained, and the time width that is most often handled is set as the width of the time condition to suggest (step 132). This process is performed for all documents (step 133).

  The time width of each document is obtained from the time information received from the search unit 120. When only one time information is included in one document, the difference between the minimum time and the maximum time of the time information is set as the document time width. Here, in the example of FIG. 6, 1920 (minimum time) to 1945 (maximum time) is the document time range, and the difference of 26 years is the document time width. When a plurality of pieces of time information are included in one document, the difference between the minimum time among the minimum times of the plurality of time information and the maximum time among the maximum times of the plurality of time information is set as the document time width. In addition, there are other methods for obtaining the time width of the document, such as a method using the time width of the time information having the highest frequency or a method using the widest time width.

  Based on the time width of the document obtained above, the width of the time condition to suggest is obtained. The width of the time condition for suggesting is the most frequent time width (step 134).

2) Document scoring A time width conformance score of each document is obtained on the basis of whether or not the time width of each document is close to the width of the time condition for suggestion (step 135). A document closer to the suggested time condition range obtained above gives a higher time width conformance score. As an example of scoring, the following formula can be considered.

ここで、score(d)は文書ｄの時間幅適合スコア、Rはサジェストする時間条件の幅、range(d)は文書ｄの時間幅を表す。

Here, score (d) represents the time width conformance score of the document d, R represents the width of the time condition to suggest, and range (d) represents the time width of the document d.

3) Extraction of time condition Suggest a time condition having the time width R obtained in the above procedure.

  First, a document time fitness score is assigned to each unit time included in the document time range (step 136). Here, the document time range is a range from the minimum time to the maximum time used when the time width of the document is obtained. Each unit time is the minimum unit of time handled by the system. For example, when the minimum unit of time handled by the system is one day, the score for a document with a time range of October 1, 2011 to October 7, 2011 and a time range conformance score of 10 is October 1, 2011. To 10 October 2011, a score of 10 will be awarded for each day.

  A document score is assigned for each unit time (step 137), and a time range having the highest score is extracted from the time range having the time width R obtained in step 134 as shown in FIG. 7 (step 138). ). The extracted time range is set as a time condition for suggestion, and is output to the time condition suggestion output unit 140 (step 139).

  Step 140) The time condition suggestion output unit 140 outputs the ranked suggestion candidates received from the time condition suggestion unit 130 and presents them to the user.

  With the above processing, it is possible to suggest to the user a time condition corresponding to the width of the seasonal period that has become a topic for each search condition (query).

  The operation of each component of the time condition presenting apparatus 100 shown in FIG. 2 can be constructed as a program and installed in a computer used as the time condition presenting apparatus to be executed or distributed via a network. It is.

  The present invention is not limited to the above-described embodiments, and various modifications and applications are possible within the scope of the claims.

100 time condition presentation device 110 query input unit 120 search unit 130 time condition suggestion unit 140 time condition suggestion output unit 150 document DB (database)

Claims (7)

A time condition presenting method for presenting a time condition in a range search,
Based on the input search condition, the search means refers to a document storage means that holds a keyword for each document and time information obtained by converting the time expression included in the document into the minimum unit handled by the system. A search step for retrieving time information;
A time condition determining step for determining a time condition based on the time information of the document;
A time condition output means for outputting the time condition;
And
In the time condition determining step,
Determining a document time width from the time information of the document, and a suggestion time width determination step as a time condition width for suggesting a time width of the most frequent document;
A document scoring step of assigning to each time unit in the time range of the document a score that increases as the time width of the document approaches the width of the time condition to suggest;
A time condition extracting step of summarizing scores for each time unit, and extracting a time range having the highest score as a time condition among time ranges having a range of time conditions to be suggested,
A time condition presenting method characterized by comprising: In the suggestion time width determination step,
2. The time condition presenting method according to claim 1, wherein a difference between the minimum time and the maximum time of the time information is a time width of the document. In the suggestion time width determination step,
2. The time condition presenting method according to claim 1, wherein a time width having the highest frequency among the time information is set as a time width of the document. A time condition presentation device for presenting a time condition in a range search,
A document storage means for holding a keyword for each document and time information obtained by converting a time expression included in the document into a minimum unit handled by the system;
Search means for searching for time information of a document by referring to the document storage means based on an input search condition;
Time condition determining means for determining a time condition based on time information of the document;
Time condition output means for outputting the time condition;
Have
The time condition determining means includes
A time width of the document is determined from the time information of the document, and a time width determination unit that determines the time width of the most frequent document as a width of a time condition for suggesting;
A document scoring means for assigning a score that is higher for a document whose time width of the document is closer to the width of the time condition to suggest to each time unit in the time range of the document;
A time condition extracting unit that counts the score for each time unit and extracts the time range having the highest score as the time condition among the time ranges having the time condition range to be suggested,
A time condition presentation device comprising: The suggestion time width determining means includes:
The time condition presenting apparatus according to claim 4, wherein the time width of the document is a difference between the minimum time and the maximum time of the time information. The suggestion time width determining means includes:
The time condition presenting apparatus according to claim 4, wherein a time width having the highest frequency among the time information is set as a time width of the document. Computer
The time condition presentation program for functioning as each means of the time condition presentation apparatus of any one of Claims 4 thru | or 6.

Images