JP2011227689A - Retrieval score calculation device and method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a retrieval score calculation device and method not to reflect unique use state by a few people to a total result regarding keywords used for retrieval.SOLUTION: A retrieval score calculation device (10) includes a retrieval log DB (111) that stores a retrieval query that a retrieval server (20) receives from a terminal (30), the time when the retrieval query is received, and terminal information that specifies the terminal that sends the retrieval query as a retrieval log after associating them. A first totalling section (142) calculates the number of terminal information on the terminal that sends the predetermined retrieval query as a first value regarding the predetermined retrieval query received by the retrieval server (20) in a first period. A second totalling section (143) calculates a second value regarding the predetermined retrieval query received by the retrieval server (20) in a second period that is different from the first period in the same way. Then a retrieval score calculation section (144) calculates a proportion of the first value and the second value as retrieval score, and an output section (145) outputs it.

Description

  The present invention relates to a search score calculation apparatus and method. In particular, the present invention relates to a search score calculation apparatus and method for calculating a score of a search query that has been used frequently within a predetermined period.

  When a user tries to investigate a topic that is attracting attention, a search site that provides a search service uses a keyword related to the topic to perform a search to obtain information. In many cases, the keywords are aggregated at a search site and ranked as a keyword having a high degree of attention. By viewing these ranking displays, the user can share information about the topic that is attracting attention with other users.

  The usage status of keywords changes depending on the time and words. For example, some keywords, such as zip codes and famous website names, have been used frequently over a long period of time, but they were introduced on TV, etc., or a big event was held. There are also keywords whose usage frequency increases rapidly.

  In view of this, with respect to the transition of the frequency of use of these keywords, an apparatus is disclosed that can check a keyword whose use frequency has suddenly increased (for example, Patent Document 1 and Patent Document 2). In each of the devices disclosed in Patent Literature 1 and Patent Literature 2, a certain keyword is tabulated based on the number of times a keyword is used for a search in a certain specific period, and the usage frequency ranking is determined.

JP 2004-013417 A JP 2004-206517 A

  However, if the value used for tabulation is the number of times a keyword is used for a search, even if a specific minority user uses a certain keyword, the keyword may be ranked higher in the ranking. In this case, the keyword cannot be a keyword that is of interest to a majority of users other than a small number of users. As described above, in the devices disclosed in Patent Document 1 and Patent Document 2, the actual use of the keyword and the ranking result may be different.

  Therefore, an object of the present invention is to provide a search score calculation apparatus and method that does not reflect the actual usage unique to a small number of people in a tabulation result for keywords used for search.

  (1) A search score calculation device for calculating a search score of a predetermined search query received from a terminal by the search server, wherein the search server receives the search query received from the terminal, the time when the search query is received, and Search log storage means for associating and storing terminal information for identifying a terminal that has transmitted a search query, and a terminal that has transmitted the predetermined search query for the predetermined search query received by the search server within a first period First counting means for counting the number of the terminal information as a first value, and the predetermined search query received by the search server within a second period different from the first period. Second counting means for counting the number of the terminal information of the terminal that transmitted the search query as a second value, and a ratio of the first value to the second value as a search score A search score calculation means for calculating, search score calculating device and an output means for outputting the search score.

  The search score calculation device according to (1) stores, as a search log, the search server associates the search query received from the terminal, the time when the search query was received, and the terminal information that identifies the terminal that transmitted the search query. . Then, the search score calculation apparatus obtains a first value by counting the number of terminal information of the terminal that has transmitted the predetermined search query for the predetermined search query in the second period relative to the first period, Similarly, a second value is obtained for a predetermined search query received by the search server within a second period different from the above period. Then, the search score calculation device calculates and outputs the ratio of the first value to the second value as the search score.

  Thereby, the search score calculation apparatus can obtain a search score indicating the frequency of use of a predetermined search query in the first period. For example, the search score calculation device can create a ranking of search queries that are frequently used in the first period by comparing the search scores for each predetermined search query. In addition, since the search score calculation apparatus obtains a search score based on the ratio of the number of terminal information of the terminals that transmitted the predetermined search query in the second period relative to the first period, the predetermined terminal performs the predetermined search. The search score can be obtained without being affected by a specific usage situation such as when a large number of queries are transmitted.

  (2) The search score calculation device according to (1), wherein the search score calculation means calculates the search score using the logarithm of each of the first value and the second value.

The search score calculation device described in (2) calculates a search score using the logarithm of each of the first value and the second value in addition to the search score calculation device described in (1).
Thereby, even if the first value and the second value are very large values, the range of possible values can be reduced and the search score can be calculated.

  (3) The search score calculation device according to (1) or (2), wherein the second period is a period before the first period.

  The search score calculation apparatus according to (3) performs a search assuming that the second period is a period earlier than the first period in addition to the search score calculation apparatus according to (1) or (2). Calculate the score. Thereby, the degree of increase from the past search score can be calculated as the search score.

  (4) The search score calculation apparatus according to (1) or (2), wherein a part of the second period overlaps with a part of the first period.

  In addition to the search score calculation device according to (1) or (2), the search score calculation device according to (4) overlaps a part of the second period with a part of the first period. Calculate the search score. Thereby, about the period which overlaps with the 2nd period and the 1st period, compared with the case where it does not overlap, the change of the number of the terminal information of the overlap period is reflected in the search score, and the overlap period It is possible to reduce the influence on the search score due to the change in the number of terminal information.

  (5) The search score calculation device according to any one of (1) to (4), further including an input unit that receives input regarding the predetermined search query and the first period.

  In addition to the search score calculation device according to any one of (1) to (4), the search score calculation device according to (5) includes an input unit that receives input regarding a predetermined search query and a first period. Prepare. Thereby, since the search score detection apparatus can accept inputs for a predetermined search query and the first period, it can obtain a search score for any search query accepted by the input means. Further, the search score calculation device can obtain the search score by the length of the first period received by the input means. For example, the search score calculation device can obtain search scores for periods (first periods) of different lengths for the search query received by the input means.

  (6) A search score calculation method for calculating a search score of a predetermined search query received from a terminal by the search server, wherein the computer receives the search query and the search query received from the terminal by the search server Search log storage means for associating and storing time and terminal information for identifying the terminal that transmitted the search query, the computer having the search server received for the predetermined search query within a first period, Counting the number of terminal information of terminals that have transmitted a predetermined search query as a first value, and the predetermined search query received by the search server within a second period different from the first period The step of counting the number of the terminal information of the terminal that has transmitted the predetermined search query as a second value, and the second of the first value Calculating a ratio value as the search score, search score calculation method to execute, and outputting the search score.

  The search score calculation method described in (6) can achieve the same effects as the search score calculation device described in (1).

  ADVANTAGE OF THE INVENTION According to this invention, the search score calculation apparatus and method which do not reflect the specific use actual condition by a few persons in a total result about the keyword used for a search can be provided.

1 is a diagram illustrating an overall configuration of a Web search system and functional blocks of a search score calculation apparatus according to an embodiment of the present invention. It is a figure which shows the search log table which concerns on embodiment of this invention. It is a flowchart which shows the flow of the process which calculates the search score which concerns on embodiment of this invention.

  Embodiments of the present invention will be described below with reference to the drawings. The embodiment described below is merely an example, and the technical scope of the present invention is not limited to this.

  A search score calculation apparatus 10 according to an embodiment of the present invention will be described with reference to FIG. FIG. 1 is a diagram illustrating an overall configuration of a Web search system 1 and functional blocks of a search score calculation apparatus 10.

  The search score calculation device 10 is connected to the communication network N and receives and stores a search query accepted by the search server 20 from the search server 20. And the search score calculation apparatus 10 calculates the search score which shows the use frequency in the search server 20 about the memorize | stored search query.

  The communication network N is formed by an optical line, a wired communication network (public line network) such as an ADSL (Asymmetric Digital Subscriber Line), or a mobile communication network having a mobile phone or PHS as a terminal. Specifically, the communication network N is configured by a computer network such as a LAN (Local Area Network) or the Internet.

  The search server 20 has a search engine 21 and provides a search service.

  The terminal 30 is a terminal used by the user. The terminal 30 is a so-called personal computer, and includes at least a CPU (Central Processing Unit) that controls the operation of the terminal 30, an input unit such as a keyboard and a mouse, and a display unit such as a display. Although there are two terminals 30 shown in FIG. 1, a larger number may be used.

  The search score calculation device 10 includes a storage unit 11 that mainly stores log information about a search query received from the search server 20 (hereinafter referred to as a search log), and an operation unit that receives an input operation to the search score calculation device 10 and the like. 12, a display unit 13 that displays processing results of various processes in the search score calculation apparatus 10, and a control unit 14 that executes various processes in the search score calculation apparatus 10.

  The storage unit 11 can be read and written by a computer using a hard disk device, a magneto-optical disk device, a nonvolatile memory such as a ROM (Read Only Memory), a volatile memory such as a RAM (Random Access Memory), or a combination thereof. It is constituted by a simple storage medium.

  The storage unit 11 includes a search log DB (Database) 111 that stores a search log received from the search server 20 by the search score calculation device 10 as a search log storage unit. The search log DB 111 stores a search log table 50.

  Here, the search log table 50 will be described with reference to FIG. FIG. 2 is a diagram showing the search log table 50.

  The search log table 50 has a “search query” column, a “reception time” column, and a “terminal IP” column, which are associated with each other. In the “search query” column, a search query received from the terminal 30 by the search server 20 is stored. The “reception time” column stores the time at which the search server 20 received the search query from the terminal 30 and received it. The “terminal IP” column stores the IP (Internet Protocol) address of the terminal that transmitted the search query to the search server 20. The terminal IP is information used as terminal information for specifying the terminal 30 that has transmitted a predetermined search query to the search server 20.

  Returning to FIG. 1, the operation unit 12 receives input to the search score calculation apparatus 10 and is configured by a keyboard, a mouse, a touch pen, a touch panel, and the like. In the present embodiment, the operation unit 12 receives input for a first period, a predetermined search query, and a second period, which are information for calculating a search score.

  The display unit 13 displays a screen for accepting data input, or displays a screen for a search score calculation result by the search score calculation device 10, and includes a cathode ray tube display device (CRT), a liquid crystal display device (LCD). ) And the like.

  The control unit 14 includes a CPU and the like, and controls the execution of each process of the search score calculation device 10. The control unit 14 may realize the function by reading a program for realizing the function of each of the following units from the storage unit 11 and executing the program, or may be realized by dedicated hardware. It may be a thing.

  The control unit 14 receives a search log from the search server 20, operates the search log acquisition unit 140 that stores the received search log in the search log DB 111, the first period, the second period, and a predetermined search query. About an input receiving unit 141 as an input unit that receives an input from the unit 12, a first totaling unit 142 as a first totaling unit for obtaining a first value in a first period for a predetermined search query, and a predetermined search query A second totaling unit 143 as a second totaling unit for obtaining a second value in the second period, and a search score calculating unit as a search score calculating unit for calculating a search score from the first value and the second value 144 and an output unit 145 as output means for outputting a search score.

  The search log acquisition unit 140 receives log information about the search query received by the search server 20 from the terminal 30 from the search server 20. Then, the search log acquisition unit 140 stores the received search log in the search log DB 111. The search log acquisition unit 140 may transmit a request for transmitting a search log to the search server 20 every predetermined period, and may receive the search log from the search server 20, or search from a predetermined database storing the search log A log may be received.

  The input receiving unit 141 receives input for the first period, the second period, and the predetermined search query via the operation unit 12. Here, the first period is a period having a predetermined length received by the input receiving unit 141. For example, the input receiving unit 141 displays an input field for each of the first period, the second period, and the predetermined search query on the display unit 13 and receives an input from the operation unit 12. Note that the input receiving unit 141 may receive an input from the terminal 30 via the communication network N.

  The 1st totalization part 142 calculates | requires the 1st value in a 1st period about a predetermined | prescribed search query. Specifically, the first tabulation unit 142 refers to the search log table 50 and counts the number of terminal IPs that have transmitted a predetermined search query to the search server 20 in the first period as a first value. At this time, the first counting unit 142 collects overlapping terminal IPs into one terminal IP and counts the number of times a predetermined search query is received. That is, the first totalization unit 142 obtains the first value based on the number of terminals 30 that have transmitted a predetermined search query in the first period to the search server 20.

  The 2nd totaling part 143 calculates | requires the 2nd value in a 2nd period about a predetermined | prescribed search query. The second period is preferably a period that does not overlap with the first period and is earlier than the first period. For example, the second period is longer than the length of the first period. The second totaling unit 143 refers to the search log table 50 and counts the number of terminal IPs that have transmitted a predetermined search query to the search server 20 within the second period as a second value. At this time, like the first totaling unit 142, the second totaling unit 143 collects overlapping terminal IPs into one terminal IP and counts the number of times a predetermined search query is received. That is, the second totalization unit 143 obtains the second value based on the number of terminals 30 that have transmitted a predetermined search query in the second period to the search server 20.

  Note that although the second period does not overlap with the first period, the present invention is not limited to this. For example, a part of the second period may overlap with a part of the first period. By doing in this way, the influence degree to the search score calculated by the search score calculation part 144 mentioned later about the period which overlaps with a 2nd period and a 1st period can be reduced. That is, when a part of the first period overlaps with a part of the second period, the first value and the second value in the overlapping period reflect the amount of change in the overlapping period. It will be. However, as the search score, the amount of change in the overlapping period is canceled out, and the influence of the amount of change on the search score can be reduced compared to the case where the amount of change does not overlap.

The search score calculation unit 144 calculates a search score using the first value and the second value. The search score calculation unit 144 calculates a search score by the following calculation formula.

  The value of this search score becomes high when the product of the first value and the second value is large. For example, it is assumed that the search scores of a plurality of search queries are calculated by setting the first period as a predetermined day and the second period as two weeks before the previous day of the predetermined day. In this case, in a plurality of search queries, a search query having a relatively high search score has a relatively larger product of the first value and the second value than other search queries. That is, this search query has a high possibility that the first value is relatively large and the second value is relatively small. Thus, a search query with a high search score can be considered as a search query that has risen sharply compared to other search queries on a given day compared to other search queries.

  Here, the normalization coefficient means the number of searches of all search queries received by the search server 20 in the second period. Thereby, the search score according to the fluctuation | variation of the search number of all the search queries can be calculated | required.

式（２）における第１の値及び第２の値は以下の式により算出してもよい。すなわち、検索スコア算出部１４４は、検索スコアを算出する際に、以下（３）、（４）、（６）及び（７）の式に示されるように、第１の値及び第２の値それぞれの対数を使用して検索スコアを算出してもよい。検索スコア算出部１４４は、第１の値及び第２の値それぞれの対数を上記式（１）に当てはめて検索スコアを算出する。これにより、それぞれの値が取り得る範囲を小さくすることができる。

Further, when calculating the search score, the search score calculation unit 144 may generalize the above formula (1) and calculate the search score by the following calculation formula.

The first value and the second value in equation (2) may be calculated by the following equations. That is, when the search score calculation unit 144 calculates the search score, the first value and the second value are expressed by the following expressions (3), (4), (6), and (7). A search score may be calculated using each logarithm. The search score calculation unit 144 calculates the search score by applying the logarithm of each of the first value and the second value to the above equation (1). Thereby, the range which each value can take can be made small.

Here, for the first value, the search score calculation unit 144 uses the first value in Expression (1), that is, the number of terminal IPs that have transmitted a predetermined search query to the search server 20 in the first period. May be. In addition, the search score calculation unit 144 uses, for the first value, a value obtained by normalizing the search query received by the search server 20 most frequently in the first period by the number of terminal IPs that transmitted the search query. May be. That is, you may use the value calculated | required by following formula (5).

In addition, the search score calculation unit 144 may use a value calculated using the following expression for the second value. For example, in the second period, a negative value may be given to the search query transmitted by the terminal 30 of the terminal IP that is half or more of the total number of the terminal IP in the second period. That is, you may use the following formula | equation (6).

Further, as the second value, a ratio between the value of the search query having the largest number of use times and the number of use times of the predetermined search query in the second period may be used. That is, the following formula (7) may be used.

As for the normalization coefficient, the sum of squares of scores for all queries of the aggregate data may be set to 1. That is, by using the following formula (8), when the first period and the second period are fixed, cosine normalization can be performed when it is desired to compare between search queries.

  The output unit 145 outputs the search score calculated by the search score calculation unit 144. In the present embodiment, the output unit 145 displays the search score on the display unit 13. Further, the output unit 145 may transmit the search score to the terminal 30 via the communication network N.

  With reference to FIG. 3, the process in which the search score calculation apparatus 10 calculates a search score is demonstrated. FIG. 3 is a flowchart showing a flow of processing in which the search score calculation apparatus 10 calculates a search score.

  In step S11, the input receiving unit 141 of the control unit 14 receives input for a predetermined search query for the first period, the second period, and the like.

  In step S12, the first totaling unit 142 of the control unit 14 totals the first value. The first totaling unit 142 refers to the search log DB 111 and counts the number of terminals 30 (number of terminal IPs) that transmitted a predetermined search query in the first period to the search server 20 as a first value.

  In step S13, the second totaling unit 143 of the control unit 14 totals the second value. The second totaling unit 143 refers to the search log DB 111 and counts the number of terminals 30 (number of terminal IPs) that transmitted a predetermined search query in the second period to the search server 20 as a second value.

  In step S14, the search score calculation unit 144 of the control unit 14 calculates a search score. The search score calculation unit 144 calculates a search score from the first value obtained by the first aggregation unit 142 and the second value obtained by the second aggregation unit 143.

  In step S15, the output unit 145 of the control unit 14 outputs the search score calculated by the search score calculation unit 144 to the display unit 13, and ends the process of this flowchart.

  According to the above-described embodiment, the search score calculation apparatus 10 includes the search log DB 111 that stores the search log together with the terminal information that identifies the terminal 30 that has transmitted the search query to the search server 20. In addition, the search score calculation apparatus 10 counts the number of terminals that have transmitted a predetermined search query to the search server 20 based on the terminal information stored in the search log DB 111 and obtains a first value by the first tabulation unit 142. And a second totaling unit 143. For this reason, for example, when the search server 20 receives a large amount of the same search query from a very small number of terminals, the search score calculation device 10 reflects the number of times the same search query is received from the same terminal 30 in the search score. Can be prevented. In such a case, it is because it cannot be said that it is a search query which many users are paying attention to. According to the present embodiment, the search score calculation device 10 can accurately obtain a search score for a search query used by a large number of users.

  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.

DESCRIPTION OF SYMBOLS 1 Search system 10 Search score calculation apparatus 11 Memory | storage part 12 Operation part 13 Display part 14 Control part 20 Search server 21 Search engine 30 Terminal 111 Search log DB
140 Search log acquisition unit 141 Input reception unit 142 First tabulation unit 143 Second tabulation unit 144 Search score calculation unit 145 Output unit N Communication network

Claims (6)

A search score calculation device for calculating a search score of a predetermined search query received by a search server from a terminal,
A search log storage unit that stores the search query received from the terminal, the time when the search query is received, and the terminal information that identifies the terminal that transmitted the search query;
First counting means for counting, as a first value, the number of terminal information of terminals that have transmitted the predetermined search query for the predetermined search query received by the search server within a first period;
For the predetermined search query received by the search server within a second period different from the first period, the number of the terminal information of the terminal that has transmitted the predetermined search query is counted as a second value. A second counting means;
Search score calculating means for calculating a ratio of the first value to the second value as a search score;
A search score calculation device comprising: output means for outputting the search score. The search score calculation device according to claim 1, wherein the search score calculation means calculates the search score using the logarithm of each of the first value and the second value. The search score calculation apparatus according to claim 1, wherein the second period is a period that is past the first period. The search score calculation device according to claim 1 or 2, wherein a part of the second period overlaps with a part of the first period. The search score calculation apparatus according to any one of claims 1 to 4, further comprising an input unit that receives an input regarding the predetermined search query and the first period. A search score calculation method in which a search server calculates a search score of a predetermined search query received from a terminal,
The computer includes a search log storage unit that stores the search query received from the terminal by the search server, the time when the search query is received, and the terminal information that identifies the terminal that transmitted the search query in association with each other.
In the computer,
For the predetermined search query received by the search server within a first period, counting the number of terminal information of the terminal that has transmitted the predetermined search query as a first value;
For the predetermined search query received by the search server within a second period different from the first period, the number of the terminal information of the terminal that has transmitted the predetermined search query is counted as a second value. Steps,
Calculating a ratio of the first value to the second value as a search score;
A search score calculation method for executing the step of outputting the search score.

Images