JP5445071B2 - Search information analysis program, search information analysis device, and search information analysis method - Google Patents

Description

  The present invention relates to a search information analysis program, a search information analysis device, and a search information analysis method, and more particularly to a search information analysis program, a search information analysis device, and a search information analysis method for analyzing a search log of a search engine.

  2. Description of the Related Art Conventionally, a service for analyzing a search log recorded in a search engine on the Internet (a Web site that provides a keyword search function) and presenting a topic word is provided (for example, Patent Document 1). . By using such a service, the user can easily know what is currently being talked about in the world, things, people or the like.

JP 2004-206517 A

  However, simply presenting a word makes it difficult for the user to understand the meaning and background of the presented word. For example, when a certain person name is presented as a topical word, it has not been understood why the person is a topic.

  The present invention has been made in view of the above points, and is a search information analysis program, a search information analysis apparatus, and a search that can appropriately support the presentation of information indicating the background of a topic word The purpose is to provide an information analysis method.

  Therefore, in order to solve the above-described problem, the search information analysis program stores the search word in the computer based on a search log in which a search word including one or more words and attribute information of the search word are stored in association with each other. Based on a relationship generation procedure for extracting a relationship between words included, associating a word related to each word and recording it in the relationship information storage means, and a topic word that is one of the words included in the search word The peripheral words that are the words having the relationship with the topic word are extracted with reference to the relationship information storage means, and the first word having the relationship with the peripheral words and the first word are extracted for each of the extracted peripheral words. A second word corresponding to the peripheral word other than the peripheral word is extracted with reference to the relation information storage means, and the number of the extracted first words and the second word Count the number of each word and the number of the first word The peripheral words calculates the degree indicating the background of the topic word based on the number of fine said second words, to execute a selection procedure for extracting a portion of the peripheral word based on the degree to which the calculated.

  According to the disclosed technology, it is possible to appropriately support the presentation of information indicating the background of a topic word.

It is a figure which shows the hardware structural example of the search information analysis apparatus in embodiment of this invention. It is a figure which shows the function structural example of the search information analysis apparatus in embodiment of this invention. It is a flowchart for demonstrating the process sequence by a search information analysis apparatus. It is a figure which shows the structural example of a search log memory | storage part. It is a figure for demonstrating the example of the extraction method of a topic word. It is a figure which shows the example of extraction of the relationship between the words based on the commonality of a jump destination URL. It is a figure which shows the example of extraction of the relationship between the words based on a search user's commonality. It is a figure which shows the example of extraction of the relationship between the words based on whether it is contained in the same search keyword. It is a flowchart for demonstrating the process sequence of the extraction process of the relationship between words based on the commonality of a jump destination URL. It is a figure which shows the example of a relationship management table. It is a flowchart for demonstrating the process sequence of the extraction process of the relationship between the words based on a search user's commonality. It is a flowchart for demonstrating the process sequence of the extraction process of the relationship between words based on whether it is contained in the same search keyword. It is a flowchart for demonstrating the process sequence of the selection process of a representative word. It is a figure which shows the example of the related information of the word within two steps from a topic word. It is a figure which shows each statistical value for every peripheral word in the specific example of related information. It is a figure which shows the example of a display of the representative word etc. regarding a topic word.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram illustrating a hardware configuration example of a search information analysis apparatus according to an embodiment of the present invention. 1 includes a drive device 100, an auxiliary storage device 102, a memory device 103, a CPU 104, an interface device 105, a display device 106, and an input connected to each other via a bus B. Device 107.

  A program for realizing processing in the search information analysis apparatus 10 is provided by a recording medium 101 such as a CD-ROM. When the recording medium 101 on which the program is recorded is set in the drive device 100, the program is installed from the recording medium 101 to the auxiliary storage device 102 via the drive device 100. However, the program need not be installed from the recording medium 101 and may be downloaded from another computer via a network. The auxiliary storage device 102 stores the installed program and also stores necessary files and data.

  The memory device 103 reads the program from the auxiliary storage device 102 and stores it when there is an instruction to start the program. The CPU 104 realizes functions related to the search information analysis device 10 according to a program stored in the memory device 103. The interface device 105 is used as an interface for connecting to a network. The display device 106 displays a GUI (Graphical User Interface) or the like by a program. The input device 107 is a keyboard, a mouse, or the like, and is used for inputting various operation instructions.

  FIG. 2 is a diagram illustrating a functional configuration example of the search information analysis apparatus according to the embodiment of the present invention. In the figure, the search information analysis apparatus 10 includes a search log storage unit 11, a topic word extraction unit 12, a relationship extraction unit 13, a representative word selection unit 14, an output processing unit 15, and the like. Each of these units is realized by processing executed by the CPU 104 by a program installed in the search information analysis apparatus 10.

  The search log storage unit 11 is a storage area for storing a search log (search history) of a search engine in the auxiliary storage device 102. A search engine is a Web site that provides a search service for information published on the Internet. The search log is data including a URL (Uniform Resource Locator) of browsed information and a search keyword (search word) used for searching the information. In other words, the search log includes the input search keyword and information (jump destination URL) indicating which information in the search result is selected as the browsing destination (jump destination or transition destination). The search log storage unit 11 may be included in another computer or storage device connected to the search information analysis apparatus 10 via a network.

  The topic word extraction unit 12 analyzes the search log recorded in the search log storage unit 11 and extracts a topic word from search keywords used for past searches. A topic word is a word that is likely to be a topic. The relationship extraction unit 13 analyzes the search log recorded in the search log storage unit 11 and extracts a relationship between words included in the search keyword based on information recorded in the search log together with the search keyword. The relationship extraction unit 13 records information indicating the extracted relationship in the auxiliary storage device 102 or the memory device 103 as a relationship management table.

  The representative word selection unit 14 selects a representative word from words included in the search keyword based on the topic word extracted by the topic word extraction unit 12 and the relationship management table generated by the relationship extraction unit 13. select. The representative word is selected based on a predetermined rule from words having a relationship with the topic word as expected to indicate the background in which the topic word is a topic. For example, it is expected that a word indicating a field or category to which a topic word belongs or a word obtained by abstracting a topic word is selected as a representative word. The output processing unit 15 outputs the representative word selected by the representative word selecting unit 14. Various output destinations such as the display device 106, the auxiliary storage device 102, or a printer (not shown) can be applied.

  Hereinafter, the processing procedure of the search information analysis apparatus 10 of FIG. 1 will be described. FIG. 3 is a flowchart for explaining the processing procedure by the search information analysis apparatus. The process shown in FIG. 6 is started by, for example, an input of a process start instruction by an operator or automatic detection of a periodic timing.

  In step S <b> 101, the topic word extraction unit 12 extracts a topic word from search keywords included in the search log recorded in the search log storage unit 11.

  FIG. 4 is a diagram illustrating a configuration example of the search log storage unit. In the figure, the search log storage unit 11 stores a date, a user identifier, a search keyword, a jump URL, and the like for each search log. The search keyword is a character string (including a single character) input as a search seed. The search keyword can include a plurality of words or character strings. In the figure, the search keyword in the search log on the second line is “Macaroon sweets”. This indicates that “Macaroon” and “Sweet” are simultaneously input as search keywords. The jump destination URL indicates the URL of the information selected as the browsing target from the search result (information list) searched based on the search keyword. For example, in the search log on the first line, information identified by “http://www.xxxxxxx.com/” is searched from information searched based on the search keyword “macaron”. It shows that.

  The date and time is the date and time when the search log was recorded, that is, the date and time when the search engine received an acquisition request for information related to the jump destination URL. The user identifier is an identifier of a user who has performed a search related to the search log. The user identifier is transmitted from the client device side (Web browser) based on a cookie corresponding to the search engine, for example. However, an identifier for each client device (for example, an IP address) may be used as the user identifier.

  FIG. 5 is a diagram for explaining an example of a topic word extraction method. The figure shows an example in which a score based on the increase rate of the number of searches over the past three days is calculated for each word included in the search keyword. Specifically, the sum of the rate of increase in the number of searches from 2 days ago to the number of searches from 3 days ago and the rate of increase in the number of searches from 2 days ago to 1 day ago is taken as the score of each word. Ranks (ranks) are determined in descending order of scores, and words having relatively high ranks (for example, words up to the top N) are set as topic words. The number of searches is the number included in the search keyword.

  In the example of the figure, the search numbers of “Macaroon” three days ago, two days ago, and one day ago are 10, 15, and 30, respectively. The increase rate from 3 days ago to 2 days ago is 1.5. The rate of increase from 2 days ago to 1 day ago is 2.0. Therefore, the score is 1.5 + 2.0 = 3.5. Since the score is the highest, “Macaron” is ranked at the top and is extracted as a topic word together with “Ramen” and “PC” in the top three.

  Other known techniques may be used for the topic word extraction process. That is, the topic word extraction method applicable to the present embodiment is not limited to a predetermined one. The topic word extraction process may be performed by a computer other than the search information analysis apparatus 10.

  Subsequently, the relationship extraction unit 13 extracts a relationship between words included in the search keyword based on the search log, and generates a relationship management table in which the extraction result is recorded (S102). The search log range for extracting the relationship between words may be the same as or different from the search log range (for example, the past three days) from which topic words are extracted.

  The relationship between words can be extracted or derived from information recorded in the search log, and is not limited to a specific relationship as long as it is reasonable. For example, in this embodiment, the relationship between words is extracted based on the commonality of jump URLs, the commonality of search execution entities (search users), or whether they are included in the same search keyword. .

  FIG. 6 is a diagram illustrating an example of extracting relationships between words based on commonality of jump destination URLs. In the example of the figure, the search keyword of the search log whose URL1 is the destination URL two days ago includes “Macaroon”, “France”, or “Tokyo”. Therefore, in this case, it is determined that “Macaroon”, “France”, and “Tokyo” are related to each other. That is, the mutual relationship between “Macaroon”, “France”, and “Tokyo” is extracted.

  Further, two days ago, the search keyword of the search log in which URL2 is the destination URL includes “Macaron”, “Italy”, “Sweets”, or “Ginza”. Therefore, in this case, the mutual relationship between “macaron”, “Italy”, “confectionery”, and “Ginza” is extracted.

  One day ago, the search keyword of the search log in which URL1 is the destination URL includes “Macaron”, “Sweets”, or “Tokyo”. Therefore, in this case, the mutual relationship between “macaron”, “sweets”, and “Tokyo” is extracted.

  Further, one day ago, the search keyword of the search log in which URL2 is the jump destination URL includes “Macaroon” or “France”. Therefore, in this case, the mutual relationship between “Macaroon” and “France” is extracted.

  When the relationships extracted in the above four cases are merged (integrated), the relationship information D1 is obtained. In the relationship information D1, the relationship between words is indicated by a line segment. The numerical value attached to the line segment indicates the number of extractions of the relationship between words connected by the line segment. For example, the relationship between “Macaroon” and “France” is extracted twice in total in the case where the destination URL of URL2 is two days ago and the case of URL2 where the destination URL is one day ago. Therefore, “2” is attached to the line segment connecting the two words.

  The jump destination URL indicates information browsed by the user. The browsed information is likely to be information intended by the user. And when the same information is intended for search, the words included in the search keyword are highly likely to have similar meanings. Therefore, the relationship extracted based on the commonality of the jump destination URLs is considered to be based on the semantic closeness (approximation degree) between words.

  Moreover, FIG. 7 is a figure which shows the example of extraction of the relationship between the words based on the commonality of a search user. In the example shown in the figure, two days ago, the user 1 performs a search by inputting a search keyword including “macaron”, “map”, or “sweets”. Therefore, in this case, it is determined that “macaroons”, “maps”, and “sweets” are related to each other. That is, the mutual relationship of “macaron”, “map”, and “sweets” is extracted.

  In addition, two days ago, the user 2 performs a search by inputting a search keyword including “macaron”, “sweets”, or “map”. Therefore, in this case, the mutual relationship between “macaron”, “sweets”, and “map” is extracted.

  Further, one day ago, the user 1 performs a search by inputting a search keyword including “macaron” or “map”. Therefore, in this case, the mutual relationship between “macaron” and “map” is extracted.

  Further, one day ago, the user 2 performs a search by inputting a search keyword including “macaron”, “France”, “Ginza”, or “recipe”. Therefore, in this case, the mutual relationship between “macaron”, “France”, “Ginza”, and “recipe” is extracted.

  When the relationships extracted in the above four cases are merged, relationship information D2 is obtained. The notation of the relationship information D2 is the same as that of the relationship information D1.

  When searching for target information, the user may input various search keywords and restart the search. In such a case, it is considered that the words included in the search keyword input have a relationship centered on the target information in the user's consciousness. Therefore, the relationship extracted based on the commonality of the search users is considered to indicate the relationship recognized by the search user.

  FIG. 8 is a diagram illustrating an example of extracting relationships between words based on whether or not they are included in the same search keyword. In the example shown in the figure, the search keyword “sweets macaroon tokyo” is entered two days ago. “Sweets Macaroon Tokyo” indicates that “Sweets”, “Macaroons”, and “Tokyo” are separated by a space or the like and input simultaneously (as one search keyword (query)). In this case, the mutual relationship of “sweets”, “macarons”, and “Tokyo” is extracted from the search keyword. The mutual relationship between the words included in the search keyword is also extracted from other search keywords two days ago and each search keyword one day ago. As a result, the relationship information D3 is obtained. The notation of the relationship information D3 is the same as that of the relationship information D1.

  Any one of the three methods shown in FIGS. 6 to 8 may be adopted, or two or more methods may be combined.

  Each processing procedure for realizing the above three methods will be described. When the extraction example of FIG. 6 is adopted, the relationship extraction unit 13 executes the processing shown in FIG.

  FIG. 9 is a flowchart for explaining a processing procedure for extracting a relationship between words based on commonality of the jump destination URL.

  In step S201, the relationship extraction unit 13 classifies search logs within a predetermined period into groups based on commonality (identity) of jump destination URLs. That is, the search logs having the same jump destination URL belong to the same group.

  Subsequently, the relationship extraction unit 13 sets one group as a processing target (S202). Hereinafter, the group to be processed is referred to as “current group”. Subsequently, the relationship extraction unit 13 registers all combinations in the case where two words are combined in the search keyword of the search log belonging to the current group, in the relationship management table (S203).

  FIG. 10 is a diagram illustrating an example of a relationship management table. The relationship management table stores the number of extractions for each combination of two words (“word 1” and “word 2”). That is, in step S203, the relationship extraction unit 13 newly registers an unregistered combination in the relationship management table, and records “1” as the number of extractions for the combination. On the other hand, for a combination that has already been registered, “1” is added to the number of extractions for the combination.

  After executing Steps S202 and S203 for all groups (Yes in S204), the relationship extraction unit 13 deletes records whose extraction count is less than the threshold from the relationship management table (S205). This is to eliminate information noise by eliminating a low-relevance relationship.

  The relationship information D1 to D3 in FIGS. 6 to 7 expresses the information recorded in the relationship management table in a visually easy-to-understand manner.

  When the extraction example of FIG. 7 is employed, the relationship extraction unit 13 executes the processing shown in FIG. FIG. 11 is a flowchart for explaining the processing procedure for extracting the relationship between words based on the commonality of search users.

  In step S211, the relationship extraction unit 13 classifies search logs within a predetermined period into groups based on the commonality (identity) of user identifiers. That is, search logs having a common user identifier belong to the same group.

  Thereafter, steps S212 to S215 are the same processing procedure as steps S202 to S205 of FIG.

  When the extraction example of FIG. 8 is employed, the relationship extraction unit 13 executes the processing shown in FIG. FIG. 12 is a flowchart for explaining a processing procedure of a relationship extraction process between words based on whether or not they are included in the same search keyword.

  In step S221, the relationship extraction unit 13 selects one search log that includes two or more search keywords in the search logs within a predetermined period. The search log selected here is hereinafter referred to as “current log”.

  Subsequently, the relationship extraction unit 13 registers all combinations in the case where two words included in the search keyword of the current log are combined in the relationship management table (S222). The format of the relationship management table may be the same as in FIG. The relationship extraction unit 13 newly registers an unregistered combination in the relationship management table, and records “1” as the number of extractions for the combination. On the other hand, for a combination that has already been registered, “1” is added to the number of extractions for the combination.

  After executing steps S221 and S222 for all search logs within a predetermined period (Yes in S223), the relationship extraction unit 13 deletes records whose extraction count is less than the threshold from the relationship management table (S224).

  Returning to FIG. Subsequent to step S102, the representative word selection unit 14 selects one of the topic words extracted in step S101 as a processing target (S103). There is no particular restriction on the order of selection of topic words. All topic words may be finally processed. Alternatively, a topic word to be processed may be selected by an instruction input from the operator. Hereinafter, the selected topic word is referred to as “current topic word”.

  Subsequently, the representative word selection unit 14 extracts records related to the relationship within two steps from the current topic word from the relationship management table (S104). Within two steps means within two branches in the relationship information D1 and the like. For example, when “macaroon” is the current topic word, “pizza” and “pasta” and the like are two steps from “macaroon” in the relationship information D3 (FIG. 8). On the other hand, “Sweets”, “France”, and “Italy” are one step from “Macaroon”. In the related information D1 to D3, words of three steps or more from “Macaroon” are not displayed for convenience.

  Extraction from the relationship management table of records indicating relationships within two steps from the current topic word may be performed as follows. First, all records including the current topic word (hereinafter referred to as “one step record”) are extracted from the relationship management table. Subsequently, a record including the word (hereinafter referred to as “two-step record”) using a word that is not the current topic word as a key in the one-step record is extracted from the records excluding the one-step record.

  Subsequently, the representative word selection unit 14 executes a representative word selection process based on the extracted records (that is, the one-step record and the two-step record) (S105). That is, a representative word is selected for each topic word. Subsequently, the output processing unit 15 outputs the selected representative word.

  When the selection and output of representative words for all topic words is completed (Yes in S107), the process in FIG. 3 ends.

  Next, details of step S105 will be described. FIG. 13 is a flowchart for explaining a processing procedure of representative word selection processing.

  In step S301, the representative word selection unit 14 extracts adjacent words of the current topic word from one step record. An adjacent word means a word within one step. Therefore, words (words other than the topic word) included in the one-step record are adjacent words of the topic word.

  For example, FIG. 14 is a diagram illustrating an example of relation information of words within two steps from a topic word. In the figure, assuming that the current topic word is “Macaroon”, in step S301, “Italy”, “Ginza”, “France”, “Sweets”, “Rollcake”, and “Sweets” are adjacent words. Extracted. The adjacent words of the topic word are particularly called “neighboring words”.

  Subsequently, the representative word selection unit 14 counts the number of adjacent words of each peripheral word based on the relationship management table (S302). Specifically, for each peripheral word, the number of records including the peripheral word in “word 1” or “word 2” is counted. For example, in FIG. 14, “Ginza” which is one of the peripheral words includes “Italy”, “France”, “Macaroon”, “Sweets”, “Sweets”, “Accessories”, and “Brand Shop”. There are seven adjacent words. Therefore, the number of adjacent words for “Ginza” is “7”.

  Subsequently, the representative word selection unit 14 counts the number of other neighboring words in the neighboring words of each neighboring word based on the relationship management table (S303). Specifically, for each peripheral word, the peripheral word is included in one of “word 1” or “word 2”, and the other word is a peripheral word other than the peripheral word (that is, adjacent to the topic word) The number of records that are word) is counted. For example, in FIG. 14, “Ginza” which is one of the peripheral words includes “Italy”, “France”, “Macaroon”, “Sweets”, “Sweets”, “Accessories”, and “Brand Shop”. There are seven adjacent words. Of these, five are "Italy", "France", "Macaroon", "Sweets", and "Sweets". Therefore, the number of neighboring words in the adjacent word for “Ginza” is “5”.

Subsequently, the representative word selection unit 14 calculates a score for each peripheral word, and selects a peripheral word having the maximum score as a representative word (S304). Here, the score is calculated by the following equation, and corresponds to an example of the degree indicating the background of the topic word in the present embodiment.
(Wn × wn) / rn
rn: Number of adjacent words (value counted in step S302)
wn: Number of neighboring words in the adjacent word (value counted in step S303)
Regarding the example of the relationship information in FIG. 14, the value of rn, the value of wn, and the value of (wn × wn) / rn for each peripheral word are shown in FIG. 15. FIG. 15 is a diagram illustrating each statistical value for each neighboring word in a specific example of the relationship information.

  In the example of the figure, the peripheral word having the maximum score is “sweets”. Therefore, “sweets” is selected as a representative word for the topic word “macaron”. In addition, not only the peripheral word having the maximum score but also a plurality of peripheral words having a score equal to or higher than a predetermined value or a score ranking higher than a predetermined value may be selected as a presentation target for showing the background of the topic word.

  When based on the example of FIG. 15, in step S106 of FIG. 3, for example, a display as shown in FIG. 16 is performed. FIG. 16 is a diagram illustrating a display example of representative words and the like related to topic words.

  In the figure, with the topic word “Macaroon” at the center, the representative word “Sweets” is placed on the upper side, and peripheral words with higher scores (“Ginza”, “France”, and “Sweets”) are placed on the lower side. Yes. The user who browses the screen shown in the figure is that “Macaroon” is a topic word and is a sweet that is a kind of sweet and is sold in Ginza. It can also be assumed that France is native. Although only representative words may be displayed, it is possible to more clearly guess the meaning, content, background, etc. of the topic word by presenting other peripheral words.

  Here, the significance of a score calculation formula ((wn × wn) / rn) used for evaluation of peripheral words will be described. wn is the number of neighboring words among neighboring words of the neighboring words. In other words, it is the number of words that are neighboring words of neighboring words and also neighboring words of topic words. Therefore, it can be said that the word is a word that has a strong relationship with peripheral words and a strong relationship with topic words. It can be considered that the larger the number of words, the stronger the relationship with the topic word (the larger the surrounding word, the greater the value of wn). Therefore, such peripheral words (neighboring words having a large value of wn) should be given high evaluation in selecting a representative word for the topic word.

  However, the value of wn is greatly influenced by the meaning range (abstraction level) of neighboring words. Specifically, a peripheral word having a wide meaning range (high abstraction) is likely to have a relationship with many words. Then, the absolute strength of the relationship between the peripheral word and the topic word, that is, the value of wn naturally increases. On the other hand, if it is a peripheral word having a narrow meaning range (low abstraction level), related words are also limited. Then, the absolute strength of the relationship between the peripheral word and the topic word, that is, the value of wn is naturally reduced. Therefore, if the strength of the relationship between each peripheral word and the topic word is simply compared by comparing wn, the strength is not properly evaluated.

  Therefore, in this embodiment, rn is used as a divisor. rn is the number of adjacent words of the peripheral word, and includes the number of words that are not adjacent to the topic word. Therefore, rn can be said to be a value indicating the total number of words closely related to the peripheral words, that is, a value indicating the meaning range of the peripheral words.

  By dividing wn by rn, the magnitude of the range of meaning of each peripheral word can be discarded, and the strength of the relationship between each peripheral word and the topic word can be compared. In other words, for each peripheral word, the ratio between each peripheral word and the topic word depends on the ratio of the relationship with the topic word (strictly the relationship with the peripheral word) with respect to all the relationships with the other words of the peripheral word. The strength of the relationship can be evaluated. Note that wn is squared because the absolute strength of the relationship with the topic word is weighted in the present embodiment. Therefore, it is not always necessary to square wn. Further, in order to change the weighting of wn and rn, each may be multiplied by a weighting coefficient. In this case, the score calculation formula is (α × wn) / (β × rn). Here, α is a weighting coefficient for wn. β is a weighting coefficient for rn.

  As described above, according to the present embodiment, the search log is analyzed to extract the relationship information between the words included in the search keyword, and the topic word extracted from the search keyword is extracted based on the relationship information. Peripheral words and representative words can be extracted or determined. Therefore, it is possible to appropriately support the presentation of information indicating the background of the topic word. That is, by presenting the extracted peripheral words and representative words, the user can easily understand the background of the topic words.

  As mentioned above, although the Example of this invention was explained in full detail, this invention is not limited to such specific embodiment, In the range of the summary of this invention described in the claim, various deformation | transformation・ Change is possible.

Regarding the above description, the following items are further disclosed.
(Appendix 1)
On the computer,
Based on a search log in which a search word including one or more words and attribute information of the search word are stored in association with each other, a relationship between words included in the search word is extracted and related to each word A relationship generation procedure for associating words with each other and storing them in the relationship information storage means;
Based on a topic word that is one of the words included in the search word, a peripheral word that is a word having the relationship with the topic word is extracted with reference to the relation information storage means, and for each extracted peripheral word The first word having the relationship with the peripheral word and the second word that is the first word and corresponds to the peripheral word other than the peripheral word are extracted with reference to the relationship information storage means And counting the number of the extracted first words and the number of the second words, respectively, so that the peripheral word is the topic word based on the number of the first words and the number of the second words. The search information analysis program which calculates the degree which shows the background of this, and performs the selection procedure which extracts a part of said peripheral word based on this calculated degree.
(Appendix 2)
The search information analysis program according to supplementary note 1, wherein the degree calculated by the selection procedure is a value obtained by dividing the number of the second words by the number of the first words.
(Appendix 3)
The attribute information of the search term is a URL of browsing destination information corresponding to the search term,
The search information analysis program according to appendix 1 or 2, wherein the relationship generation procedure extracts a relationship between the words based on a commonality of URLs of the browsing destination information.
(Appendix 4)
The attribute information of the search term is an identifier of a search subject corresponding to the search term,
The search information analysis program according to appendix 1 or 2, wherein the relation generation procedure extracts a relation between the words based on commonality of the identifiers of the search subjects.
(Appendix 5)
The search information analysis program according to supplementary note 1 or 2, wherein the relationship generation procedure extracts a relationship between the words based on whether or not they are included in the same search word in the search log.
(Appendix 6)
Based on a search log in which a search word including one or more words and attribute information of the search word are stored in association with each other, a relationship between words included in the search word is extracted and related to each word Relationship generation means for associating words with each other and storing them in the relationship information storage means;
Based on a topic word that is one of the words included in the search word, a peripheral word that is a word having the relationship with the topic word is extracted with reference to the relation information storage means, and for each extracted peripheral word The first word having the relationship with the peripheral word and the second word that is the first word and corresponds to the peripheral word other than the peripheral word are extracted with reference to the relationship information storage means And counting the number of the extracted first words and the number of the second words, respectively, so that the peripheral word is the topic word based on the number of the first words and the number of the second words. A search information analysis apparatus comprising: a selection unit that calculates a degree indicating the background of the word and extracts a part of the peripheral words based on the calculated degree.
(Appendix 7)
The search information analysis apparatus according to appendix 6, wherein the degree of calculation by the selection means is a value obtained by dividing the number of the second words by the number of the first words.
(Appendix 8)
The attribute information of the search term is a URL of browsing destination information corresponding to the search term,
The search information analysis apparatus according to appendix 6 or 7, wherein the relationship generation means extracts a relationship between the words based on a commonality of URLs of the browsing destination information.
(Appendix 9)
The attribute information of the search term is an identifier of a search subject corresponding to the search term,
The search information analysis device according to appendix 6 or 7, wherein the relationship generation means extracts the relationship between the words based on the commonality of the identifiers of the search subjects.
(Appendix 10)
The search information analysis apparatus according to appendix 6 or 7, wherein the relationship generation means extracts a relationship between the words based on whether or not the search log includes the same search word.
(Appendix 11)
Computer
Based on a search log in which a search word including one or more words and attribute information of the search word are stored in association with each other, a relationship between words included in the search word is extracted and related to each word A relationship generation procedure for associating words with each other and storing them in the relationship information storage means;
Based on a topic word that is one of the words included in the search word, a peripheral word that is a word having the relationship with the topic word is extracted with reference to the relation information storage means, and for each extracted peripheral word The first word having the relationship with the peripheral word and the second word that is the first word and corresponds to the peripheral word other than the peripheral word are extracted with reference to the relationship information storage means And counting the number of the extracted first words and the number of the second words, respectively, so that the peripheral word is the topic word based on the number of the first words and the number of the second words. And a selection procedure for extracting a part of the neighboring words based on the calculated degree and calculating a degree indicating the background of the search information.
(Appendix 12)
The search information analysis method according to supplementary note 11, wherein the degree calculated by the selection procedure is a value obtained by dividing the number of the second words by the number of the first words.
(Appendix 13)
The attribute information of the search term is a URL of browsing destination information corresponding to the search term,
The search information analysis method according to appendix 11 or 12, wherein the relation generation procedure extracts a relation between the words based on a commonality of URLs of the browsing destination information.
(Appendix 14)
The attribute information of the search term is an identifier of a search subject corresponding to the search term,
The search information analysis method according to appendix 11 or 12, wherein the relationship generation procedure extracts a relationship between the words based on commonality of the identifiers of the search subjects.
(Appendix 15)
The search information analysis method according to appendix 11 or 12, wherein the relationship generation procedure extracts a relationship between the words based on whether or not the search log is included in the same search word.

DESCRIPTION OF SYMBOLS 10 Search information analysis apparatus 11 Search log memory | storage part 12 Topic word extraction part 13 Relation extraction part 14 Representative word selection part 15 Output processing part 100 Drive apparatus 101 Recording medium 102 Auxiliary storage apparatus 103 Memory apparatus 104 CPU
105 interface device 106 display device 107 input device B bus

Claims (7)

On the computer,
Based on a search log in which a search word including one or more words and attribute information of the search word are stored in association with each other, a relationship between words included in the search word is extracted and related to each word A relationship generation procedure for associating words with each other and storing them in the relationship information storage means;
Based on the topic word is one of the words included in the retrieval word, the surrounding words is the word with the relationship between the topic words extracted by referring to the relation storing means, each extracted peripheral Language to a first word having a relationship to the peripheral words, extracts the second word corresponding to peripheral word and the extracted one of the first word by referring to the relation storing means, said The number of the extracted first words and the number of the second words are counted, respectively, and based on the number of the first words and the number of the second words, the surrounding words indicate the background of the topic word. A search information analysis program that calculates a degree to be shown and executes a selection procedure for extracting a part of the peripheral words based on the calculated degree.   The search information analysis program according to claim 1, wherein the degree calculated by the selection procedure is a value obtained by dividing the number of the second words by the number of the first words. The attribute information of the search term is a URL of browsing destination information corresponding to the search term,
The search information analysis program according to claim 1, wherein the relationship generation procedure extracts a relationship between the words based on a commonality of URLs of the browsing destination information. The attribute information of the search term is an identifier of a search subject corresponding to the search term,
The search information analysis program according to claim 1 or 2, wherein the relationship generation procedure extracts a relationship between the words based on a commonality of identifiers of the search subjects.   The search information analysis program according to claim 1 or 2, wherein the relationship generation procedure extracts a relationship between the words based on whether or not the search log is included in the same search word. Based on a search log in which a search word including one or more words and attribute information of the search word are stored in association with each other, a relationship between words included in the search word is extracted and related to each word Relationship generation means for associating words with each other and storing them in the relationship information storage means;
Based on the topic word is one of the words included in the retrieval word, the surrounding words is the word with the relationship between the topic words extracted by referring to the relation storing means, each extracted peripheral Language to a first word having a relationship to the peripheral words, extracts the second word corresponding to peripheral word and the extracted one of the first word by referring to the relation storing means, said The number of the extracted first words and the number of the second words are counted, respectively, and based on the number of the first words and the number of the second words, the surrounding words indicate the background of the topic word. A search information analysis apparatus comprising: a selection unit that calculates a degree to be shown and extracts a part of the peripheral words based on the calculated degree. Computer
Based on a search log in which a search word including one or more words and attribute information of the search word are stored in association with each other, a relationship between words included in the search word is extracted and related to each word A relationship generation procedure for associating words with each other and storing them in the relationship information storage means;
Based on the topic word is one of the words included in the retrieval word, the surrounding words is the word with the relationship between the topic words extracted by referring to the relation storing means, each extracted peripheral Language to a first word having a relationship to the peripheral words, extracts the second word corresponding to peripheral word and the extracted one of the first word by referring to the relation storing means, said The number of the extracted first words and the number of the second words are counted, respectively, and based on the number of the first words and the number of the second words, the surrounding words indicate the background of the topic word. A search information analysis method for calculating a degree to be shown and performing a selection procedure for extracting a part of the peripheral words based on the calculated degree.

Images