JP5136181B2 - Information retrieval device - Google Patents

Description

  The present invention relates to an information search apparatus that searches for information, and more particularly to an information search apparatus that can appropriately exclude information unnecessary for a user from a search result.

  In recent years, a device (hereinafter referred to as an information recording device) that records and plays back a TV program (or content distributed over the Internet) such as a PC (Personal Computer) equipped with a TV function or a PVR (Personal Video Recorder). Is widely spread.

  And this information recording device is provided with various functions in order to improve the convenience of the user, for example, using an electronic program guide (metadata) distributed from EPG (Electric Program Guide), It has a function of automatically searching and presenting a user's favorite program and a function of automatically recording the searched program.

  However, there are cases where programs searched by the information recording apparatus include programs that are unnecessary for the user, such as re-broadcasts of dramas and movies that have already been viewed by other broadcast stations.

  Therefore, in Patent Document 1, the program recorded by the user in the past is stored as an operation history, the search result is compared with the operation history, and the program recorded by the user in the past is excluded from the search result. A technique for removing a program unnecessary for a user from a search result is disclosed.

  Specifically, in Patent Document 1, a combination of a program genre, a title, and a broadcasting station included in the operation history is specified, and a program having the same combination as the specified combination is removed from the search result. When the genre of the program is a movie, a program with the same title is removed from the search result.

Japanese Patent No. 3747847

  However, the above-described conventional technique has a problem that it is not possible to appropriately set an exclusion condition for removing a program unnecessary for a user from a search result.

  FIG. 23 is a diagram for explaining the problems of the prior art. FIG. 23A shows a user operation history, and FIG. 23B shows a search result before an unnecessary program is removed. FIG. 23C shows search results after unnecessary programs are removed by the conventional technique.

  Here, the program with the program identification number “1011” shown in FIG. 23C is unnecessary because it is a rebroadcast of the program with the program identification number “1001” shown in the operation history (FIG. 23A). Although it should be removed as a program, it cannot be removed from the search result because the combination of the broadcasting station and the title is different.

  Similarly, the program with the program identification number “1013” shown in (c) of FIG. 23 is unnecessary because it is a rebroadcast of the program with the program identification number “1002” shown in the operation history ((a) of FIG. 23). Although it should be removed as a program, it cannot be removed from the search result because the combination of the broadcasting station and the title is different.

  In addition, the program with the program identification number “1012” shown in FIG. 23C is simultaneously broadcast on the program with the program identification number “1002” shown in the operation history (FIG. 23A). Therefore, it should be removed as an unnecessary program. However, when both are compared, the genre is different, so it cannot be removed from the search result.

  On the other hand, the programs with the program identification numbers “1014” and “1015” shown in FIG. 23B are the next day of the program identification numbers “1001” and “1002” shown in the operation history (FIG. 23A), respectively. It is a program that will be broadcast later, both of which are different from the operation history program, so it should be left in the search results, but when comparing both, the combination of broadcast station, genre, title matches, Accidentally excluded from search results.

  As described above, in the conventional technique, the exclusion condition for removing the unnecessary program from the search result cannot be set appropriately. Therefore, the program unnecessary for the user is removed, and the information necessary for the user is included in the search result. I can't leave it.

  The present invention has been made in order to solve the above-described problems caused by the prior art, and provides an information search apparatus capable of appropriately setting exclusion conditions so that a user can recommend a necessary program. With the goal.

  In order to solve the above-described problems and achieve the object, this information search device includes a metadata acquisition unit that acquires and stores metadata of information to be searched, contents of past operation history by the user, and operations An operation history recording unit that stores operation history information associated with metadata of target information, metadata of search target information stored by the metadata acquisition unit, and storage by the operation history recording unit The index creation unit that creates each index from the text part of the metadata of the information to be operated and the respective indexes created by the index creation unit to obtain each search result, and then the operation history information And the index similarity between the search result information or the index similarity between the search result information is within a predetermined range, and A search unit that extracts a search result that satisfies a predetermined condition for a comparison result other than the text portion of the data as an excluded search result that is a search result of a candidate to be excluded, and stores information on the exclusion search result extracted by the recommendation unit Based on the exclusion history recording unit, the information on the exclusion search result stored in the exclusion history recording unit, and the operation history information stored in the operation history recording unit, the recommendation unit performs the exclusion search result And an exclusion condition adjustment unit that adjusts the conditions for extracting.

  According to this information search device, a condition for extracting an exclusion search result based on information on an exclusion search result stored in the exclusion history recording unit and information on an operation history stored in the operation history recording unit Therefore, such conditions can be appropriately set according to the user's needs, and information unnecessary by the user can be excluded from the search results.

  Exemplary embodiments of an information search apparatus according to the present invention will be explained below in detail with reference to the accompanying drawings.

  First, an outline and features of the information search apparatus according to the first embodiment will be described. The information search apparatus according to the first embodiment compares an index generated from program metadata (such as electronic program guide data) with an index generated based on the user's preference, and matches the user's preference. The search result of the program considered to be once extracted. Then, the information search apparatus excludes a program unnecessary for the user from the search result using a predetermined exclusion condition, information on the excluded program (hereinafter, excluded data), and a history of operations on the user's program ( The exclusion condition is adjusted based on the operation history data.

  As described above, since the information search apparatus according to the first embodiment adjusts the exclusion condition based on the exclusion data and the operation history data, the exclusion condition can be appropriately set and only the program unnecessary for the user is searched. Can be removed from the results.

  Next, the configuration of the information search apparatus 100 according to the first embodiment will be described. FIG. 1 is a functional block diagram of the configuration of the information search apparatus 100 according to the first embodiment. As shown in the figure, the information retrieval apparatus 100 includes a metadata acquisition unit 110, an operation history recording unit 120, a preference information setting unit 130, an index creation unit 140, a recommendation unit 150, and an exclusion history recording unit. 160, an exclusion condition adjustment unit 170, a display unit 180, and a search result recording unit 190. The information search apparatus 100 includes a communication apparatus connected to a network and a receiving apparatus that receives broadcast waves (not shown).

  Among these, the metadata acquisition unit 110 is means for acquiring and storing metadata such as a search target program, a net moving image, and news (news distributed on the net). FIG. 2 is a diagram illustrating an example of a data structure of metadata stored in the metadata acquisition unit 110. As shown in the figure, this metadata includes a program identification number for identifying a program, a broadcast date and time, a broadcast station, a broadcast length, a broadcast wave, a genre, a title, and a description of the program. Have.

  Note that the metadata in FIG. 2 indicates the metadata of the program. Net video metadata includes video distribution start date, distribution length, distribution source, genre, title, description, distribution deadline, etc. News metadata includes news distribution start date, distribution source, genre, title, Consists of explanation, delivery deadline, etc.

  The metadata acquisition unit 110 acquires program metadata from an EPG (Electric Program Guide) superimposed on a broadcast wave or an EPG provided through a network. Further, the metadata acquisition unit 110 acquires the metadata of the net moving image and the news from the distribution source of the net moving image and the distribution source of the news.

  The operation history recording unit 120 is a means for recording various operation contents (viewing, reproduction, browsing, recording, automatic reproduction, saving, etc.) performed by the user for a program, an online video, news, etc., as operation history data. is there.

  FIG. 3 is a diagram illustrating an example of the data structure of the operation history data according to the first embodiment. As shown in the figure, the operation history data includes a program identification number, an operation date and time, an operation that is operation content, a broadcast date and time, a broadcasting station, a genre, a title, and a description. The operation history recording unit 120 deletes an entry from the operation history data when the number of entries recorded in the operation history data reaches a specified number or when an entry whose recording period is longer than the specified period is deleted. Good.

  The preference information setting unit 130 is a means for setting a search condition for searching for a user's favorite program, network video, news, and the like. The preference information setting unit 130 may acquire keywords from the user via an input device (not shown), set each acquired keyword as a search condition, or operation history data recorded in the operation history recording unit 120. May be acquired and set as a search condition.

  The index creation unit 140 includes an index of metadata (search target) of the metadata acquisition unit 110 (hereinafter referred to as search target index), an index of operation history data of the operation history recording unit 120 (hereinafter referred to as operation history index), and preference information setting. This is means for creating an index of search conditions (hereinafter referred to as a search condition index) of the unit 130. The index creation unit 140 outputs the created search target index, operation history index, and search condition index information to the recommendation unit 150.

  The index is created by dividing the text included in the title, description, genre, etc. into keywords through morphological analysis or n-gram analysis, and then weighting them using statistics such as the appearance frequency of each keyword. ). Hereinafter, a method for creating a search target index, an operation history index, and a search condition index will be described in order.

(About search target index)
A method of creating a search target index will be described using the metadata program identification number “1001” shown in FIG. 2 as an example. When the text included in the title, description, and genre of the program identification number “1001” is decomposed into keywords by performing morphological analysis, the index creating unit 140 includes the keywords “Uptown, Music,...”. In this case, the basis of the program identification number “1001” in the vector space is “Uptown”, “Music”,.

  The size of each base is a statistic of each keyword. The statistics are calculated using, for example, TF (Term Frequency) or DF (Document Frequency). TF is the frequency at which a specific keyword appears in the text of a certain program (including online video and news), and DF is included in all the programs to be searched (for example, included in the metadata shown in FIG. 2). The number of programs including a specific keyword in all programs).

The specific formula for calculating the statistic (base size) is
Statistics = TF / DF
It becomes.

  For example, the frequencies (TF) at which the keywords “Uptown” and “Music” appear in the text of the program identification number “1001” shown in FIG. 2 are “0.8” and “0.6”, respectively. If the numbers (DF) of the programs including the keywords “Uptown” and “Music” are “2” and “3”, respectively, the program identification number “1001” is shown. The index is “Uptown (base size 0.4), Music (base size 0.2)”. Note that the search target indexes are created for the number of programs included in the metadata shown in FIG.

(About the operation history index)
A method for creating an operation history index will be described using the search history data “1001” shown in FIG. 3 as an example. When the text included in the title, description, and genre of the program identification number “1001” is decomposed into keywords by performing morphological analysis, the index creating unit 140 includes the keywords “Uptown, Music,...”. In this case, the basis of the program identification number “1001” in the vector space is “Uptown”, “Music”,.

  The size of each base is a statistic of each keyword. The statistic calculation method is the same as the calculation method described in the search index. The frequencies (TF) at which the keywords “Uptown” and “Music” appear in the text of the program identification number “1001” shown in FIG. 3 are “0.8” and “0.6”, respectively. If the number of programs (DF) including the keywords “Uptown” and “Music” is “2” and “3”, respectively, the index of the program identification number “1001” is , “Uptown (base size 0.4), music (base size 0.2)”. Note that as many operation history indexes as the number of programs included in the operation history data shown in FIG. 3 are created.

(About search condition index)
Here, a method of creating a search condition index when using operation history data (see FIG. 3) will be described. The index creation unit 140 performs morphological analysis on all the texts included in the operation history index, and decomposes the texts into keywords. If the keyword “Uptown, Music, Dustpan,...” Is included as a result of dividing the text into keywords, the search condition index in the vector space is defined as “Uptown”, “Music”, “ Dustpan "...

  The size of each base is a statistic of each keyword. The statistic calculation method is the same as the calculation method described in the search index. The frequencies (TF) at which the keywords “Uptown”, “Music”, and “Dustpan” appear in the text of the operation history data are “0.4”, “0.3”, and “0.2”, respectively. When the number of programs (DF) including the keywords “Uptown”, “Music”, and “Dust” is “2”, “3”, and “2”, respectively, The search condition index is “uptown (base size 0.2), music (base size 0.1), dustpan (base size 0.1)”.

  The recommendation unit 150 is a unit that acquires index information (search target index, operation history index, search condition index) from the index generation unit 140, and generates a search result of a program recommended to the user based on the acquired index. is there. The recommendation unit 150 sequentially executes a process of creating a primary search result as a search result candidate and a process of removing unnecessary programs from the primary search result. In the following, a process for creating a primary search result and a process for removing unnecessary programs will be described in order.

(About processing to create primary search results)
The recommendation unit 150 calculates the similarity between each search target index and the search condition index, and creates a primary search result based on the calculated similarity. Note that the similarity between vectors is represented by an inner product (or angle, correlation coefficient) between the vectors. FIG. 4 is a diagram illustrating a method of calculating similarity between indexes. The degree of similarity increases as the information matches the user's preference.

  The recommendation unit 150 calculates the inner product of each search target index and the search condition index, and extracts a set of the search target index and the search condition index in which the calculated inner product value is equal to or greater than the first specified value (extracted) The search target index included in the set is referred to as a search result candidate index.) Various information (program identification number, broadcast date / time, broadcast station, genre, title, description) corresponding to the search result candidate index is registered in the primary search result.

(About processing to remove unnecessary programs)
The recommendation unit 150 excludes unnecessary entries from the primary search result based on the search result candidate index and the operation history index. A final search result is obtained by removing unnecessary entries from the primary search result. The recommendation unit 150 first calculates the inner product of the search result candidate index and the operation history index.

  Then, the recommendation unit 150 compares the calculated inner product with the second specified value, and sets the search result candidate index and the operation history index (hereinafter referred to as the first history index) that the inner product is equal to or greater than the second specified value. Group). The program of the search result candidate index included in the first group is a program that is extremely likely to be rebroadcast or simulcast of the paired operation history index program.

  Based on the broadcast date and time (broadcast start date, broadcast start time) of the program corresponding to each index included in the first group, the recommendation unit 150 does not require the search result candidate index included in the first group for the user. It is determined whether or not the index is a correct program.

  Specifically, the recommendation unit 150 compares the broadcast start time of the program corresponding to the search result candidate index with the broadcast start time of the program corresponding to the operation history index in the first set. If the broadcast start times are different, it is determined that the program of the search result candidate index in the first set is a rebroadcast program of a program that has already been operated by the user (program of operation history data). To do.

  In the first group, the recommendation unit 150 compares the broadcast start date of the program corresponding to the search result candidate index with the broadcast start date of the program corresponding to the operation history index. When the broadcast start dates are the same, it is determined that the program of the search result candidate index in the first set is a simulcast of a program that has already been operated by the user (program of operation history data).

  The recommendation unit 150 excludes the entry corresponding to the search result candidate index of the program determined to be rebroadcast or simulcast from the primary search result, and registers the excluded entry in the exclusion data. FIG. 5 is a diagram illustrating an example of exclusion data according to the first embodiment. As shown in the figure, the exclusion data includes a program identification number of a program excluded from the primary search result, a broadcast date and time, a broadcast station, a genre, a title, and a description.

  Subsequently, the recommendation unit 150 calculates an inner product between the search result candidate indexes (excluding a search result candidate index corresponding to a program determined to be a rebroadcast program or a simulcast program), and the calculated inner product is the second. A set of search result candidate indexes (hereinafter referred to as a second set) that is equal to or greater than the threshold is extracted.

  Based on the broadcast date and time (broadcast start date, broadcast start time) of the program corresponding to each index included in the second set, the recommendation unit 150 does not require the search result candidate index included in the second set for the user. It is determined whether or not the index is a correct program.

  Specifically, the recommendation unit 150 compares the broadcast start time of the program corresponding to one search result candidate index with the broadcast start time of the program corresponding to the other search result candidate index in the second set. . When the broadcast start times are different, it is determined that the program of one search result candidate index in the second set is a rebroadcast of the program of the other search result candidate index.

  Further, the recommendation unit 150 compares the broadcast start date of the program corresponding to one search result candidate index with the broadcast start date of the program corresponding to the other search result candidate index in the second set. When the broadcast start dates are the same, it is determined that the program of one search result candidate index in the second set is a simulcast of the program of the other search result candidate index.

  The recommendation unit 150 selects one of the search result candidate indexes in the second set including the search result candidate index of the program determined to be rebroadcast or simulcast. Then, the entry corresponding to the selected search result candidate index is excluded from the primary search result, and the excluded entry is registered in the exclusion data (see FIG. 5). The recommendation unit 150 records the created exclusion data in the exclusion history recording unit 160.

  Incidentally, the recommendation unit 150 calculates the inner product of the search result index and the operation history index, and calculates the inner product, the second specified value, and the third specified value (where the second specified value> the third specified value). Value). Then, a set of search result candidate indexes and operation history indexes (hereinafter referred to as a third set) in which the inner product is not less than the third specified value and less than the second specified value is extracted. The program of the search result candidate index included in the third set is a program that is suspected of being rebroadcasted or a simulcast of the paired operation history index. However, the possibility of rebroadcasting or simulcasting is lower than in the first group.

  The recommendation unit 150 does not require a search result candidate index to be set in the third set based on the broadcast date and time (broadcast start date, broadcast start time) of the program corresponding to each index included in the third set for the user. It is determined whether or not there is a possibility that the program index is correct.

  Specifically, the recommendation unit 150 compares the broadcast start time of the program corresponding to the search result candidate index with the broadcast start time of the program corresponding to the operation history index in the third set. If the broadcast start times are different, the program of the search result candidate index in the third group may be a rebroadcast program of a program already operated by the user (program of operation history data). Is determined to be high.

  In the third set, the recommendation unit 150 compares the broadcast start date of the program corresponding to the search result candidate index with the broadcast start date of the program corresponding to the operation history index. If the broadcast start dates are the same, the program of the search result candidate index in the third set is likely to be a simulcast of a program that has already been operated by the user (program of operation history data). Is determined.

  The recommendation unit 150 registers various information (identification number, broadcast date / time, broadcast station, genre, title, description) of the search result candidate index of the program determined to be highly likely to be rebroadcast or simulcast in the exclusion candidate data. To do. FIG. 6 is a diagram illustrating an example of the data structure of the exclusion candidate data according to the first embodiment. As shown in the figure, this exclusion candidate data includes a program identification number, a broadcast date, a broadcast station, a genre, a title, and a description. The recommendation unit 150 records the created exclusion candidate data in the exclusion history recording unit 160.

  The recommendation unit 150 outputs the remaining program entries obtained by removing unnecessary programs from the primary search result to the display unit 180 and the search result recording unit 190 as search result data. FIG. 7 is a diagram illustrating an example of a data structure of search result data according to the first embodiment.

  As shown in FIG. 7, the search result data includes a search order, a program identification number, a broadcast date and time, a broadcast station, a genre, a title, and a description. For example, the recommendation unit 150 may rearrange the entries of the search result data by setting a lower search order for an entry having a higher similarity between the search target index and the search condition index.

  The exclusion history recording unit 160 is means for storing exclusion data (see FIG. 5) and exclusion candidate data (see FIG. 6). The exclusion history recording unit 160 is used when a new search is performed, when the number of entries registered in the exclusion data or exclusion candidate data reaches a prescribed number, or when the period after recording is longer than the prescribed period. You may delete the entry which became long from exclusion data and exclusion candidate data.

  The exclusion condition adjustment unit 170 compares the exclusion data and the exclusion candidate data recorded in the exclusion history recording unit 160 with the operation history data recorded in the operation history recording unit 120, and the second condition used by the recommendation unit 150. Means for adjusting the specified value and the third specified value.

  Specifically, the exclusion condition adjustment unit 170 compares each entry included in the exclusion data with each entry of the operation history data to determine whether there is a set of entries having the same broadcast date / time, broadcast station, title, and description. Determine whether or not. When the exclusion condition adjustment unit 170 determines that there is a set of matching entries, the exclusion condition adjustment unit 170 adds a predetermined value to the second specified value used by the recommendation unit 150.

  If each entry included in the excluded data is compared with each entry in the operation history data, and there is a matching entry, it is considered that the excluded entry originally wanted to be left in the search result. Therefore, the exclusion condition is tightened by adding a predetermined value to the second specified value (the entry included in the primary search result is not easily excluded).

  On the other hand, the exclusion condition adjustment unit 170 compares each entry included in the exclusion candidate data with each entry of the operation history data to determine whether there is a set of entries having the same broadcast date / time, broadcast station, title, and description. Determine whether. If the exclusion condition adjustment unit 170 determines that there is no matching entry pair, the exclusion condition adjustment unit 170 subtracts a predetermined value from the second specified value and the third specified value used by the recommendation unit 150.

  Comparing each entry included in the exclusion candidate data with each entry in the operation history data, the fact that there is no matching entry means that the entry included in the exclusion candidate data (the entry included in the exclusion candidate data is the primary search) (Remaining in the result) is considered to be something that was originally not desired to be included in the search result, so that the exclusion condition is relaxed by subtracting a predetermined value from the second specified value and the third specified value. (Easy to exclude entries included in primary search results)

  Thus, the exclusion condition adjustment unit 170 compares the exclusion data and exclusion candidate data recorded in the exclusion history recording unit 160 with the operation history data recorded in the operation history recording unit 120, and the recommendation unit 150 Since the second specified value and the third specified value to be used are adjusted, it is possible to appropriately set search conditions for obtaining a search result that suits the user's preference.

  The display unit 180 is means for displaying the search result data (see FIG. 7) on a screen such as a display. The user refers to the screen and sets recording of the program. The search result recording unit 190 is means for storing search result data. The user refers to the search result data recorded in the search result recording unit 190 and specifies an entry to be saved among the search result data.

  Next, various processing procedures of the information search apparatus 100 according to the first embodiment will be described. FIG. 8 and FIG. 9 are flowcharts showing the process in which the exclusion condition adjustment unit 170 adjusts the second and third specified values. FIGS. 10 and 11 show the search result data by the recommendation unit 150 according to the first embodiment. It is a flowchart which shows the process which produces.

  As shown in FIG. 8, the exclusion condition adjustment unit 170 sets the value of i to 0 (step S101), acquires operation history data from the operation history recording unit 120 (step S102), and from the exclusion history recording unit 160. A list of excluded data (M) is acquired (step S103).

  Then, the exclusion condition adjustment unit 170 determines whether or not the value of i is smaller than the value of M. If the value of i is equal to or greater than the value of M (No in step S104), the process ends. . On the other hand, when the value of i is smaller than the value of M (step S104, Yes), it is determined whether the operation history data and the exclusion data (i-th) are the same information (step S105).

  If the operation history data and the excluded data (i-th) are different information (No in step S106), a value obtained by adding 1 to i is set as a new value of i (step S107), and the process proceeds to step S104. Transition. On the other hand, when the operation history data and the exclusion data (i-th) are the same information (step S106, Yes), the second specified value is adjusted (the second specified value is increased) (step S108), End the process.

  As described above, the exclusion condition adjustment unit 170 increases the second specified value when the operation history data and the exclusion data include the same entry (information). Therefore, the exclusion condition adjustment unit 170 sets the exclusion condition according to the user's preference. Can be adjusted.

  As illustrated in FIG. 9, the exclusion condition adjustment unit 170 sets the value of i to 0 (step S201), acquires operation history data from the operation history recording unit 120 (step S202), and from the exclusion history recording unit 160. A list (M) of exclusion candidate data is acquired (step S203).

  Then, the exclusion condition adjustment unit 170 determines whether or not the value of i is smaller than the value of M. If the value of i is smaller than the value of M (Yes in step S204), the operation history data and It is determined whether or not the exclusion candidate data (i-th) is the same information (step S205).

  When the operation history data and the exclusion candidate data (i-th) are the same information (step S206, Yes), the detection flag is set to ON (step S207), and a value obtained by adding 1 to i is set as a new i. (Step S208), and the process proceeds to step S204. On the other hand, when the operation history data and the exclusion candidate data (i-th) are different (No at Step S206), the process proceeds to Step S208.

  By the way, when the value of i is greater than or equal to the value of M (step S204, No), it is determined whether or not the detection flag is on (step S209), and when the detection flag is off (step S209). (S210, No), the second and third specified values are adjusted (lower the second and third specified values) (step S211), and the process is terminated. On the other hand, when the detection flag is on (step S210, Yes), the process is ended as it is.

  In this way, the exclusion condition adjustment unit 170 lowers the second and third specified values when the operation history data and the exclusion candidate data do not include the same entry, so that the exclusion condition is set according to the user's preference. Can be adjusted.

  As shown in FIG. 10, the recommendation unit 150 sets the value of i to 0 (step S301), and acquires search result candidate indexes (M) (step S302). Then, it is determined whether or not the value of i is smaller than the value of M. If the value of i is equal to or greater than the value of M (No at Step S303), the process is terminated.

  On the other hand, when the value of i is smaller than the value of M (step S303, Yes), the value of j is set to 0 (step S304), and the operation history index (N pieces) is acquired (step S305). Then, it is determined whether or not the value of j is smaller than the value of N. If the value of j is greater than or equal to the value of N (step S306, No), the process proceeds to step S317.

  On the other hand, when the value of j is smaller than the value of N (step S306, Yes), the similarity between the search result candidate index (i-th) and the operation history index (j-th) is calculated (step S307), It is determined whether or not the similarity is greater than or equal to a second specified value (step S308).

  When the similarity is less than the second specified value (No at Step S309), the process proceeds to Step S315. On the other hand, if the similarity is equal to or higher than the second specified value (step S309, Yes), it is determined whether or not the broadcast start time is different between the search result candidate index (i-th) and the operation history index (j-th). Determination is made (step S310).

  If the broadcast start times are different (step S311, Yes), the program entry corresponding to the search result candidate index (i-th) is excluded from the primary search result (step S312), and the process proceeds to step S317.

  On the other hand, if the broadcast start times are the same (step S311, No), it is determined whether the search start candidate index (i-th) and the operation history index (j-th) have the same broadcast start date (step). S313).

  If the broadcast start date is different (No in step S314), a value obtained by adding 1 to the value of j is set as a new value of j (step S315), and the process proceeds to step S306. On the other hand, if the broadcast start date is the same (step S314, Yes), the entry of the program corresponding to the search result candidate index (i-th) is excluded from the primary search result (step S316), and the value of i is set to 1. The added value is set as a new value of i (step S317), and the process proceeds to step S303.

  As described above, the recommendation unit 150 calculates the inner product of the search result candidate index and the operation history index, compares it with the second specified value, and then performs the primary search based on the broadcast start time and the broadcast start date and time of the program. Since unnecessary data is excluded from the results, a search result suitable for the user's preference can be created.

  As shown in FIG. 11, the recommendation unit 150 sets the value of i to 0 (step S401), and acquires search result candidate indexes (M) (step S402). Then, it is determined whether or not the value of i is smaller than the value of M. If the value of i is equal to or greater than the value of M (No in step S403), the process is terminated.

  On the other hand, when the value of i is smaller than the value of M (step S403, Yes), a value obtained by adding 1 to the value of i is set as the value of j (step S404), and the value of j is the value of M. If the value of j is greater than or equal to the value of M (step S405, No), a value obtained by adding 1 to the value of i is set as a new value of i (step S405). S406), the process proceeds to step S403.

  On the other hand, when the value of j is smaller than the value of M (step S405, Yes), the similarity between the search result candidate index (i-th) and the search result candidate index (j-th) is calculated (step S407). Then, it is determined whether or not the similarity is greater than or equal to a second specified value (step S408).

  When the similarity is less than the second specified value (No at Step S409), the process proceeds to Step S416. On the other hand, if the similarity is equal to or greater than the second specified value (step S409, Yes), whether or not the broadcast start time is different between the search result candidate index (i-th) and the search result candidate index (j-th). Is determined (step S410).

  If the broadcast disclosure times are different (step S411, Yes), the entry of the program corresponding to one of the search result candidate index (i-th) or the search result candidate index (j-th) is set as the primary search result. (Step S412), and the process proceeds to step S416.

  On the other hand, if the broadcast start times are the same (step S411, No), it is determined whether the search start candidate index (i-th) and the search result candidate index (j-th) have the same broadcast start date. (Step S413).

  When the broadcast disclosure date is different (step S414, No), the process proceeds to step S416. On the other hand, if the broadcast disclosure dates are the same (step S414, Yes), the program entry corresponding to one of the search result candidate index (i-th) or the search result candidate index (j-th) is selected. Then, it is excluded from the primary search result (step S415), a value obtained by adding 1 to the value of j is set as a new value of j (step S416), and the process proceeds to step S405.

  As described above, the recommendation unit 150 calculates the inner product between the search result candidate indexes and compares it with the second specified value, and then, based on the broadcast start time and the broadcast start date / time of the program, the recommendation unit 150 eliminates unnecessary from the primary search result. Since the data is excluded, a search result that suits the user's preference can be created.

  As described above, in the information search apparatus 100 according to the first embodiment, the exclusion condition adjustment unit 170 uses the operation history data of the operation history recording unit 120, the exclusion data and the exclusion candidate data of the exclusion history recording unit 160. Based on the second specified value, the recommendation unit 150 adjusts the second specified value, and provides the search result to the user by excluding unnecessary program entries from the primary search result. The exclusion condition can be appropriately set according to the user's preference, and program information meeting the user's needs can be provided.

  In addition, since the information search apparatus 100 according to the first embodiment creates a search result based on the similarity of each index and the second specified value, even when there is notation fluctuation in metadata or the like, the user Programs that are unnecessary or duplicated can be excluded from the search results.

  Further, in the information search apparatus 100 according to the first embodiment, the recommendation unit 150 excludes the program entry from the primary search result based on the broadcast start time and broadcast disclosure date and time in addition to the similarity. The search result can be provided to the user in a state in which the program that has already been operated by the user has been removed or the program to be a simulcast has been removed.

  In the first embodiment, unnecessary program entries are excluded from the primary search result based on the broadcast start date and broadcast start time of the program. However, the present invention is not limited to this. For example, when the search target is a net moving image, unnecessary moving image information can be excluded from the primary search result based on the distribution length of the net moving image. In the second embodiment, an information search apparatus 200 that deletes unnecessary video information from the primary search result based on the distribution length of the net video will be described.

  FIG. 12 is a functional block diagram of the configuration of the information search apparatus 200 according to the second embodiment. As shown in the figure, the information search apparatus 200 includes a metadata acquisition unit 210, an operation history recording unit 220, a preference information setting unit 230, an index creation unit 240, a recommendation unit 250, and an exclusion history recording unit. 260, an exclusion condition adjustment unit 270, a display unit 280, and a search result recording unit 290. The information search apparatus 200 includes a communication apparatus connected to a network and a receiving apparatus that receives broadcast waves (not shown).

  The metadata acquisition unit 210 is means for acquiring the metadata of the net moving image from the distribution source of the net moving image. The metadata of a net moving image has a moving image distribution start date and time, a distribution length, a distribution source, a genre, a title, a description, a distribution deadline, and the like.

  The operation history recording unit 220 is a means for recording various operation contents (use, reproduction, browsing, recording, automatic reproduction, saving, etc.) performed by the user on the net moving image as operation history data. FIG. 13 is a diagram illustrating an example of the data structure of the operation history data according to the second embodiment. As shown in the figure, the operation history data includes a content identification number for identifying a net video, an operation date / time, an operation, a distribution date / time, a distribution source, a distribution length, a genre, a title, a description, Have

  The preference information setting unit 230 is a means for setting a search condition for searching for a user's favorite net moving image. The preference information setting unit 230 may acquire keywords from the user via an input device (not shown), set each acquired keyword as a search condition, or operation history data recorded in the operation history recording unit 220. May be acquired and set as a search condition.

  The index creation unit 240 includes an metadata (search target) index (hereinafter referred to as a search target index) of the metadata acquisition unit 210, an operation history data index (hereinafter referred to as an operation history index) of the operation history recording unit 220, and preference information setting. This is a means for creating a search condition index (hereinafter referred to as a search condition index) of the unit 230. The index creation unit 240 outputs the created search target index, operation history index, and search condition index information to the recommendation unit 250. The index creation method is the same as that in the first embodiment, and a description thereof will be omitted.

  The recommendation unit 250 obtains index information (search target index, operation history index, search condition index) from the index creation unit 240, and creates a search result of the net video recommended to the user based on the obtained index. It is. The recommendation unit 250 sequentially executes a process of creating a primary search result as a search result candidate and a process of removing unnecessary net moving images from the primary search result. In the following, a process for creating a primary search result and a process for removing unnecessary net moving images will be described in order.

(About processing to create primary search results)
The recommendation unit 250 calculates the similarity between each search target index and the search condition index, and creates a primary search result based on the calculated similarity. Note that the similarity between vectors is represented by an inner product (or angle, correlation coefficient) between the vectors.

  The recommendation unit 250 calculates the inner product of each search target index and the search condition index, and extracts a set of the search target index and the search condition index whose calculated inner product value is equal to or greater than the first specified value (Example) Similarly to 1, the search target index included in the extracted set is referred to as a search result candidate index. Various information (content identification number, delivery date, delivery source, delivery length, genre, title, description) corresponding to the search result candidate index is registered in the primary search result.

(About the process of removing unnecessary internet video)
The recommendation unit 250 excludes unnecessary entries from the primary search result based on the search result candidate index and the operation history index. A final search result is obtained by removing unnecessary entries from the primary search result. The recommendation unit 250 first calculates the inner product of the search result candidate index and the operation history index.

  Then, the recommendation unit 250 compares the calculated inner product with the second specified value, and sets the search result candidate index and the operation history index (hereinafter referred to as the first history index) that the inner product is equal to or greater than the second specified value. Group). The net moving image of the search result candidate index included in the first set is extremely likely to be the same net moving image as the net moving image of the paired operation history index.

  The recommendation unit 250 determines whether the search result candidate index included in the first set is an index of the net moving image unnecessary for the user based on the distribution length of the net moving image corresponding to each index included in the first set. Determine whether or not.

  Specifically, the recommendation unit 250 compares the distribution length of the net moving image corresponding to the search result candidate index with the distribution length of the net moving image corresponding to the operation history index in the first set. If the difference between the distribution lengths is within the fourth specified value, the net video of the search result candidate index in the first set is the net video already operated by the user (net video of the operation history data). It is determined that it is the same online video.

  The recommendation unit 250 excludes the entry corresponding to the search result candidate index of the net moving image determined to be the same net moving image from the primary search result, and registers the excluded entry in the excluded data. FIG. 14 is a diagram illustrating an example of the data structure of the exclusion data according to the second embodiment. As shown in the figure, this exclusion data includes a content identification number, distribution date, distribution source, distribution length, genre, title, and description of the net moving image excluded from the primary search result.

  Subsequently, the recommendation unit 250 calculates the inner product between the search result candidate indexes (excluding the search result candidate index of the net moving image determined to be the same net moving image as the net moving image already operated by the user), and calculates the calculated inner product. A set of search result candidate indexes (hereinafter referred to as the second set) in which is equal to or greater than the second threshold is extracted.

  The recommendation unit 250 determines whether the search result candidate index included in the second set is an index of the net moving image unnecessary for the user based on the distribution length of the net moving image corresponding to each index included in the second set. Determine whether or not.

  Specifically, in the second set, the recommendation unit 250 compares the distribution length of the net video corresponding to one search result candidate index with the distribution length of the net video corresponding to the other search result candidate index. . When the difference between the distribution lengths is within the fourth specified value, the net video of one search result candidate index in the second set is the same net video as the net video of the other search result candidate index. Judge that there is.

  The recommendation unit 250 selects one of the search result candidate indexes in the second set including the search result candidate index of the net moving image determined to be the same net moving image. Then, the entry corresponding to the selected search result candidate index is excluded from the primary search result, and the excluded entry is registered in the exclusion data (see FIG. 14). The recommendation unit 250 records the created exclusion data in the exclusion history recording unit 260.

  Incidentally, the recommendation unit 250 calculates the inner product of the search result index and the operation history index, and calculates the inner product, the second specified value, and the third specified value (where the second specified value> the third specified value). Value). Then, a set of search result candidate indexes and operation history indexes (hereinafter referred to as a third set) in which the inner product is not less than the third specified value and less than the second specified value is extracted. The net moving image of the search result candidate index included in the third set is suspected to be the same net moving image as the net moving image of the paired operation history index.

  Based on the distribution length of the net video corresponding to each index included in the third set, the recommendation unit 250 may search the candidate search result index included in the third set as an index of net video unnecessary for the user. It is determined whether there is sex.

  Specifically, the recommendation unit 250 compares the distribution length of the net moving image corresponding to the search result candidate index with the distribution length corresponding to the operation history index in the third set. When the difference between the distribution lengths is within the fourth specified value, the net video of the search result candidate index in the third set is the net video already operated by the user (net video of the operation history data). It is determined that there is a high possibility that it is the same online video.

  The recommendation unit 250 excludes various types of information (content identification number, distribution date / time, distribution source, distribution length, genre, title, description) of the search result candidate index of the net video that is determined to be highly likely to be the same net video. Register as candidate data. The recommendation unit 250 records the created exclusion candidate data in the exclusion history recording unit 260.

  The recommendation unit 250 outputs the remaining net moving image entries obtained by removing unnecessary net moving images from the primary search result to the display unit 280 and the search result recording unit 290 as search result data. FIG. 15 is a diagram illustrating an example of a data structure of search result data according to the second embodiment.

  As shown in FIG. 15, this search result data includes a search order, a content identification number, a delivery date, a delivery source, a delivery length, a genre, a title, and a description. For example, the recommendation unit 250 may rearrange the entries of the search result data by setting a lower search order for an entry having a higher similarity between the search target index and the search condition index.

  The exclusion history recording unit 260 is means for storing exclusion data and exclusion candidate data. The exclusion history recording unit 260 excludes entries whose exclusion data or exclusion candidate data has reached the specified number, or entries whose recording period has become longer than the specified period. It may be deleted from the data.

  The exclusion condition adjustment unit 270 compares the exclusion data and the exclusion candidate data recorded in the exclusion history recording unit 260 with the operation history data recorded in the operation history recording unit 220, and uses the second data used by the recommendation unit 250. Means for adjusting the specified value and the third specified value. Since the specific processing content of the exclusion condition adjustment unit 270 is the same as that of the exclusion condition adjustment unit 170 shown in the first embodiment, the description thereof is omitted.

  The display unit 280 is means for displaying the search result data (see FIG. 15) on a screen such as a display. The user refers to such a screen and sets recording of a net moving image. The search result recording unit 290 is means for storing search result data. The user refers to the search result data recorded in the search result recording unit 290 and designates an entry to be saved among the search result data.

  Next, a processing procedure of the information search apparatus 200 according to the second embodiment will be described. FIG. 16 is a flowchart of a process in which the recommendation unit 250 according to the second embodiment creates search result data.

  As illustrated in FIG. 16, the recommendation unit 250 sets the value of i to 0 (step S501), and acquires search result candidate indexes (M) (step S502). Then, it is determined whether or not the value of i is smaller than the value of M. If the value of i is equal to or greater than the value of M (step S503, No), the process ends.

  On the other hand, when the value of i is smaller than the value of M (step S503, Yes), the value of j is set to 0 (step S504), and the operation history index (N pieces) is acquired (step S505). Then, it is determined whether or not the value of j is smaller than the value of N. If the value of j is greater than or equal to the value of N (step S506, No), the process proceeds to step S503.

  On the other hand, when the value of j is smaller than the value of N (step S506, Yes), the similarity between the search result candidate index (i-th) and the operation history index (j-th) is calculated (step S507). It is determined whether or not the similarity is greater than or equal to a second specified value (step S508).

  When the similarity is less than the second specified value (No at Step S509), the process proceeds to Step S512. On the other hand, when the degree of similarity is equal to or greater than the second specified value (step S509, Yes), the difference in distribution length between the search result candidate index (i-th) and the operation history index (j-th) is the fourth specified value. It is determined whether it is within the value (step S510).

  If the difference in delivery length is outside the fourth specified value (step S511, No), a value obtained by adding 1 to the value of j is set as a new value of j (step S512), and the process proceeds to step S506. On the other hand, if the difference in distribution length is within the fourth specified value (step S511, Yes), the entry of the net video corresponding to the search result candidate index (i-th) is excluded from the primary search result (step S513). The process proceeds to step S503.

  As described above, in the information search apparatus 200 according to the second embodiment, the recommendation unit 250 calculates the inner product of the search result candidate index and the operation history index and compares it with the second specified value. Since unnecessary data is excluded from the primary search result based on the distribution length of the moving image, it is possible to create a search result that meets the user's needs.

  In the first embodiment, unnecessary program entries are excluded from the primary search result based on the broadcast start date and broadcast start time of the program. However, the present invention is not limited to this. For example, when the search target is news distributed on the net, unnecessary news information can be excluded from the primary search result based on the distribution date of news. In the third embodiment, an information search apparatus 300 that excludes unnecessary news information from the primary search result based on the distribution date of news will be described.

  FIG. 17 is a functional block diagram of the configuration of the information search apparatus 300 according to the third embodiment. As shown in the figure, the information search apparatus 300 includes a metadata acquisition unit 310, an operation history recording unit 320, a preference information setting unit 330, an index creation unit 340, a recommendation unit 350, and an exclusion history recording unit. 360, an exclusion condition adjustment unit 370, a display unit 380, and a search result recording unit 390. The information search device 300 includes a communication device connected to a network and a receiving device that receives broadcast waves (not shown).

  The metadata acquisition unit 310 is means for acquiring news metadata from a news distribution source. News metadata includes news delivery start date and time, delivery source, genre, title, description, delivery deadline, and the like.

  The operation history recording unit 320 is a means for recording various operation contents (use, reproduction, browsing, recording, automatic reproduction, saving, etc.) performed on the news video by the user as operation history data. FIG. 18 is a diagram illustrating an example of the data structure of the operation history data according to the third embodiment. As shown in the figure, the operation history data includes a content identification number for identifying news, an operation date / time, an operation, a distribution date / time, a distribution source, a genre, a title, and a description.

  The preference information setting unit 330 is a means for setting search conditions for searching for user-favorite news. The preference information setting unit 330 may acquire keywords from the user via an input device (not shown), set each acquired keyword as a search condition, or operation history data recorded in the operation history recording unit 320 May be acquired and set as a search condition.

  The index creation unit 340 includes metadata (search target) index (hereinafter referred to as search target index) of the metadata acquisition unit 310, operation history data index (hereinafter referred to as operation history index) of the operation history recording unit 320, and preference information setting. This is a means for creating a search condition index (hereinafter referred to as a search condition index) of the unit 330. The index creation unit 340 outputs information about the created search target index, operation history index, and search condition index to the recommendation unit 350. The index creation method is the same as that in the first embodiment, and a description thereof will be omitted.

  The recommendation unit 350 is a means for acquiring index information (search target index, operation history index, search condition index) from the index generation unit 340 and generating search results of news recommended to the user based on the acquired index. is there. The recommendation unit 350 sequentially executes a process for creating a primary search result as a search result candidate and a process for removing unnecessary news from the primary search result. In the following, a process for creating a primary search result and a process for removing unnecessary news will be described in order.

(About processing to create primary search results)
The recommendation unit 350 calculates the similarity between each search target index and the search condition index, and creates a primary search result based on the calculated similarity. Note that the similarity between vectors is represented by an inner product (or angle, correlation coefficient) between the vectors.

  The recommendation unit 350 calculates the inner product of each search target index and the search condition index, and extracts a set of the search target index and the search condition index whose calculated inner product value is equal to or greater than the first specified value (Example) Similarly to 1, the search target index included in the extracted set is referred to as a search result candidate index. Then, various information (content identification number, delivery date, delivery source, genre, title, description) corresponding to the search result candidate index is registered in the primary search result.

(About processing to remove unnecessary news)
The recommendation unit 350 excludes unnecessary entries from the primary search result based on the search result candidate index and the operation history index. A final search result is obtained by removing unnecessary entries from the primary search result. The recommendation unit 350 first calculates the inner product of the search result candidate index and the operation history index.

  Then, the recommendation unit 350 compares the calculated inner product with the second specified value, and sets the search result candidate index and the operation history index (hereinafter referred to as the first history index) that the inner product is equal to or greater than the second specified value. Group). The news of the search result candidate index included in the first group is very likely to be the same as the news of the paired operation history index.

  The recommendation unit 350 determines whether or not the search result candidate index included in the first group is an index of news unnecessary for the user based on the distribution date and time of the news corresponding to each index included in the first group. Determine.

  Specifically, in the first set, the recommendation unit 350 compares the news delivery date corresponding to the search result candidate index with the news delivery date corresponding to the operation history index. When the difference between the delivery dates and times is within the fifth specified value, the news of the search result candidate index in the first set is the same as the news already operated by the user (news of operation history data). It is determined that

  The recommendation unit 350 excludes the entry corresponding to the search result candidate index of the news determined to be the same news from the primary search result, and registers the excluded entry in the exclusion data. FIG. 19 is a diagram illustrating an example of the data structure of the exclusion data according to the third embodiment. As shown in the figure, this exclusion data includes the content identification number of the news excluded from the primary search result, the distribution date and time, the distribution source, the genre, the title, and the description.

  Subsequently, the recommendation unit 350 calculates the inner product between the search result candidate indexes (excluding the search result candidate index of the news determined to be the same news as the news that the user has already operated), and the calculated inner product is the second A set of search result candidate indexes (hereinafter referred to as a second set) that is equal to or greater than the threshold is extracted.

  The recommendation unit 350 determines whether or not the search result candidate index included in the second set is an index of news unnecessary for the user based on the distribution date and time of the news corresponding to each index included in the second set. Determine.

  Specifically, in the second set, the recommendation unit 350 compares the news delivery date corresponding to one search result candidate index with the news delivery date corresponding to the other search result candidate index. When the difference between the delivery dates is within the fifth specified value, it is determined that the news of one search result candidate index in the second set is the same news as the news of the other search result candidate index To do.

  The recommendation unit 350 selects one of the search result candidate indexes in the second set including the search result candidate index of the news determined to be the same news. Then, the entry corresponding to the selected search result candidate index is excluded from the primary search result, and the excluded entry is registered in the exclusion data (see FIG. 19). Further, the recommendation unit 350 records the created exclusion data in the exclusion history recording unit 360.

  Incidentally, the recommendation unit 350 calculates the inner product of the search result index and the operation history index, and calculates the inner product, the second specified value, and the third specified value (where the second specified value> the third specified value). Value). Then, a set of search result candidate indexes and operation history indexes (hereinafter referred to as a third set) in which the inner product is not less than the third specified value and less than the second specified value is extracted. The news of the search result candidate index included in the third group is suspected to be the same news as the news of the paired operation history index.

  The recommendation unit 350 may determine that the search result candidate index included in the third group is an index of news unnecessary for the user based on the distribution date and time of the news corresponding to each index included in the third group. It is determined whether or not there is.

  Specifically, the recommendation unit 350 compares the distribution date and time of news corresponding to the search result candidate index with the distribution date and time corresponding to the operation history index in the third group. When the difference between the delivery dates and times is within the fifth specified value, the news of the search result candidate index in the third set is the same as the news already operated by the user (news of operation history data). It is determined that there is a high possibility.

  The recommendation unit 350 registers various information (content identification number, distribution date, distribution source, genre, title, description) of the news search result candidate index determined to be highly likely to be the same news in the exclusion candidate data. . The recommendation unit 350 records the created exclusion candidate data in the exclusion history recording unit 360.

  The recommendation unit 350 outputs the remaining news entries obtained by removing unnecessary news from the primary search result to the display unit 380 and the search result recording unit 390 as search result data. FIG. 20 is a diagram illustrating an example of a data structure of search result data according to the third embodiment.

  As shown in FIG. 20, the search result data includes a search order, a content identification number, a delivery date / time, a delivery source, a genre, a title, and a description. For example, the recommendation unit 350 may rearrange the entries of the search result data by setting a lower search order for an entry having a higher similarity between the search target index and the search condition index.

  The exclusion history recording unit 360 is means for storing exclusion data and exclusion candidate data. The exclusion history recording unit 360 excludes the entry whose exclusion data or exclusion candidate data has reached the specified number or the entry whose recording period is longer than the specified period from the exclusion data and the exclusion candidate. It may be deleted from the data.

  The exclusion condition adjustment unit 370 compares the exclusion data and exclusion candidate data recorded in the exclusion history recording unit 360 with the operation history data recorded in the operation history recording unit 320, and uses the second data used by the recommendation unit 350. Means for adjusting the specified value and the third specified value. Since the specific processing contents of the exclusion condition adjustment unit 370 are the same as those of the exclusion condition adjustment unit 170 shown in the first embodiment, the description thereof is omitted.

  The display unit 380 is means for displaying the search result data (see FIG. 20) on a screen such as a display. The user refers to this screen and sets news recording or the like. The search result recording unit 390 is means for storing search result data. The user refers to the search result data recorded in the search result recording unit 390, and specifies an entry to be saved among the search result data.

  Next, a processing procedure of the information search apparatus 300 according to the third embodiment will be described. FIG. 21 is a flowchart of the process in which the recommendation unit 350 according to the third embodiment creates search result data.

  As shown in FIG. 21, the recommendation unit 350 sets the value of i to 0 (step S601) and acquires search result candidate indexes (M) (step S602). Then, it is determined whether or not the value of i is smaller than the value of M. If the value of i is equal to or greater than the value of M (step S603, No), the process ends.

  On the other hand, when the value of i is smaller than the value of M (step S603, Yes), the value of j is set to 0 (step S604), and the operation history index (N pieces) is acquired (step S605). Then, it is determined whether or not the value of j is smaller than N. If the value of j is equal to or greater than the value of N (step S606, No), the process proceeds to step S603.

  On the other hand, when the value of j is smaller than the value of N (step S606, Yes), the similarity between the search result candidate index (i-th) and the operation history index (j-th) is calculated (step S607). It is determined whether or not the similarity is greater than or equal to a second specified value (step S608).

  When the similarity is less than the second specified value (No at Step S609), the process proceeds to Step S612. On the other hand, if the similarity is equal to or greater than the second specified value (step S609, Yes), the difference in delivery date and time between the search result candidate index (i-th) and the operation history index (j-th) is the fifth specified value. It is determined whether the value is within the range (step S610).

  If the difference in delivery date / time is outside the fifth specified value (step S611, No), a value obtained by adding 1 to the value of j is set as a new value of j (step S612), and the process proceeds to step S606. On the other hand, if the difference in distribution date and time is within the fifth specified value (step S611, Yes), the news entry corresponding to the search result candidate index (i-th) is excluded from the primary search result (step S613), The process proceeds to step S603.

  As described above, in the information search apparatus 300 according to the third embodiment, after the recommendation unit 350 calculates the inner product of the search result candidate index and the operation history index and compares it with the second specified value, Since unnecessary data is excluded from the primary search result based on the distribution date and time, a search result that meets the user's needs can be created.

  By the way, among the processes described in the present embodiment, all or a part of the processes described as being automatically performed can be manually performed, or the processes described as being performed manually can be performed. All or a part can be automatically performed by a known method. In addition, the processing procedure, control procedure, specific name, and information including various data and parameters shown in the above-described document and drawings can be arbitrarily changed unless otherwise specified.

  Each component of the information search apparatuses 100, 200, and 300 shown in FIGS. 1, 12, and 17 is functionally conceptual, and does not necessarily need to be physically configured as illustrated. In other words, the specific form of distribution / integration of each device is not limited to that shown in the figure, and all or a part thereof may be functionally or physically distributed or arbitrarily distributed in arbitrary units according to various loads or usage conditions. Can be integrated and configured. Furthermore, each processing function performed by each device may be realized by a CPU and a program that is analyzed and executed by the CPU, or may be realized as hardware by wired logic.

  FIG. 22 is a diagram illustrating a hardware configuration of a computer 400 that configures the information search apparatuses 100, 200, and 300 according to the first to third embodiments. As shown in FIG. 22, the computer (information retrieval device) 400 includes an input device 410, an output device 420, a RAM (Random Access Memory) 430, a ROM (Read Only Memory) 440, and a medium reader that reads data from a storage medium 450, a communication device 460 that transmits / receives data to / from other devices, a CPU (Central Processing Unit) 470, and an HDD (Hard Disk Drive) 480 are connected by a bus 490.

  The HDD 480 stores a search result creation program 480a that exhibits the same functions as the functions of the information search apparatuses 100, 200, and 300 described above. When the CPU 470 reads and executes the search result creation program 480a, the search result creation process 470a is activated. Here, the search result creation process 470a includes the metadata acquisition units 110, 210, and 310, the operation history recording units 120, 220, and 320, and the preference information setting units 130 and 230, shown in FIGS. 330, index creation unit 140, 240, 340, recommendation unit 150, 250, 350, exclusion history recording unit 160, 260, 360, exclusion condition adjustment unit 170, 270, 370, search result recording unit 190, 290, 390, etc. Correspond.

  The RAM 430 stores various data 430a used for the search result creation process 470a. The various data 430a corresponds to metadata (see FIG. 2), operation history data (see FIG. 3), exclusion data (see FIG. 5), exclusion candidate data (see FIG. 6), and the like. The CPU 470 reads out various data 430a stored in the RAM 430 and executes processing to create search result data.

  By the way, the search result creation program 480a shown in FIG. 22 is not necessarily stored in the HDD 480 from the beginning. For example, a “portable physical medium” such as a flexible disk (FD), a CD-ROM, a DVD disk, a magneto-optical disk, or an IC card inserted into a computer, or a hard disk drive (HDD) provided inside or outside the computer. The search result creation program 480a is stored in the “fixed physical medium” of “Other computer (or server)” connected to the computer via a public line, the Internet, a LAN, a WAN, or the like. The computer may read out and execute the search result creation program 480a from these.

  The following appendices are further disclosed with respect to the embodiments including the first to third embodiments.

(Supplementary Note 1) A metadata acquisition unit that acquires and stores metadata of information to be searched;
An operation history recording unit for storing operation history information in which the contents of the past operation history by the user and the metadata of the information to be operated are associated;
An index creation unit that creates each index from the metadata of the information to be searched stored in the metadata acquisition unit and the text part of the metadata of the operation target information stored in the operation history recording unit;
After obtaining each search result by comparing each index created by the index creation unit, the similarity between the index of the operation history information and the search result information, or the index of the search result information A recommendation unit that extracts a search result whose similarity is within a predetermined range and a comparison result other than the text portion of the metadata satisfies a predetermined condition as an excluded search result that is a candidate search result to be excluded;
An exclusion history recording unit for storing information of the exclusion search result extracted by the recommendation unit;
Based on the information on the exclusion search result stored in the exclusion history recording unit and the information on the operation history stored in the operation history recording unit, the condition for the recommendation unit to extract the exclusion search result is as follows. An exclusion condition adjustment section to be adjusted;
An information retrieval apparatus comprising:

(Additional remark 2) The said recommendation part is compared and searched when the similarity of the index of the information of the said search result is in a predetermined range, and comparison results other than the text part of metadata satisfy | fill predetermined conditions. The information search device according to appendix 1, wherein one of the results is extracted as an excluded search result.

(Additional remark 3) The said exclusion condition adjustment part compares the information of the said exclusion search result, and the information of the operation object preserve | saved at the said operation history recording part, and the said index is contained when the information which corresponds is contained. 3. The information search device according to appendix 1 or 2, wherein a predetermined range to be compared with the degree of similarity is narrowed.

(Additional remark 4) The said exclusion condition adjustment part produces | generates the 1st exclusion information from which a similarity becomes more than predetermined value, and the 2nd exclusion information from which a similarity becomes less than predetermined value, The said 2nd exclusion information, A supplementary note that compares the information of the operation target stored in the operation history recording unit and expands a predetermined range to be compared with the similarity of the index when no matching information is included The information search device according to 1, 2, or 3.

(Additional remark 5) The said recommendation part compares the broadcast start time of the metadata of the said operation history information and the information of a search result, or the broadcast start time of the metadata of search results, and broadcast start time differs. The information search device according to any one of appendices 1 to 4, wherein the search result is extracted as the excluded search result.

(Additional remark 6) The said recommendation part compares the broadcast start date of the metadata with the information of the said operation history, and the information of a search result, or the broadcast start date of the metadata of search results, and a broadcast start date is the same The information search device according to any one of appendices 1 to 5, wherein the search result is extracted as the exclusion search result.

(Supplementary Note 7) The recommendation unit compares the content distribution length between the operation history information and the search result information or the content distribution length between the search results, and the difference in distribution length is less than a predetermined value. The information search device according to any one of appendices 1 to 6, wherein the search result is extracted as the exclusion search result.

(Supplementary Note 8) The recommendation unit compares the content distribution date and time between the operation history information and the search result information or the content distribution date and time between the search results, and the difference in distribution date and time is less than a predetermined value. The information search device according to any one of appendices 1 to 7, wherein the search result is extracted as the exclusion search result.

(Supplementary note 9) The information retrieval device is
Obtaining metadata of information to be searched and storing it in a storage device;
Storing information of operation history in which the content of the past operation history by the user is associated with the metadata of the information to be operated in a storage device;
Creating each index from the metadata of the information to be searched stored in the storage device and the text portion of the metadata of the information to be operated;
After each of the created indexes is compared to obtain each search result, the index similarity between the operation history information and the search result information, or the index similarity between the search result information is a predetermined value. Extracting a search result within a range and a comparison result other than the text portion of the metadata that satisfies a predetermined condition as an excluded search result that is a candidate search result to be excluded;
Storing information of the excluded search result in a storage device;
Adjusting a condition for extracting the exclusion search result based on the information of the exclusion search result stored in the storage device and the information of the operation history;
The search method characterized by including.

(Appendix 10)
A metadata acquisition procedure for acquiring metadata of information to be searched and storing it in a storage device;
An operation history recording procedure for storing, in a storage device, operation history information in which the contents of past operation history by the user and the metadata of the information to be operated are associated;
An index creation procedure for creating respective indexes from the metadata of the search target information stored in the storage device and the text portion of the metadata of the operation target information;
After each of the created indexes is compared to obtain each search result, the index similarity between the operation history information and the search result information, or the index similarity between the search result information is a predetermined value. A recommendation procedure for extracting a search result that satisfies a predetermined condition within a range and a comparison result other than the text portion of the metadata as a search result of a candidate to be excluded,
Exclusion candidate recording procedure for storing information of the exclusion search result in a storage device;
An exclusion condition adjustment procedure for adjusting a condition for extracting the exclusion search result based on the information of the exclusion search result stored in the storage device and the information of the operation history;
A search program characterized in that is executed.

1 is a functional block diagram illustrating a configuration of an information search apparatus according to a first embodiment. It is a figure which shows an example of the data structure of the metadata which a metadata acquisition part stores. FIG. 3 is a diagram illustrating an example of a data structure of operation history data according to the first embodiment. It is a figure which shows the calculation method of the similarity between indexes. It is a figure which shows an example of the exclusion data concerning the present Example 1. It is a figure which shows an example of the data structure of the exclusion candidate data concerning the present Example 1. FIG. It is a figure which shows an example of the data structure of the search result data concerning the present Example 1. FIG. It is a flowchart which shows the process which the exclusion condition adjustment part concerning the present Example 1 adjusts the 2nd, 3rd prescribed value (1). It is a flowchart which shows the process in which the exclusion condition adjustment part concerning the present Example 1 adjusts the 2nd, 3rd prescribed value (2). It is a flowchart which shows the process in which the recommendation part concerning the present Example 1 produces search result data (1). It is a flowchart which shows the process in which the recommendation part concerning the present Example 1 produces search result data (2). It is a functional block diagram which shows the structure of the information search device concerning the present Example 2. It is a figure which shows an example of the data structure of the operation history data concerning the present Example 2. It is a figure which shows an example of the data structure of the exclusion data concerning the present Example 2. It is a figure which shows an example of the data structure of the search result data concerning the present Example 2. 10 is a flowchart illustrating a process in which a recommendation unit according to the second embodiment creates search result data. FIG. 10 is a functional block diagram illustrating a configuration of an information search apparatus according to a third embodiment. It is a figure which shows an example of the data structure of the operation history data concerning the present Example 3. It is a figure which shows an example of the data structure of the exclusion data concerning the present Example 3. It is a figure which shows an example of the data structure of the search result data concerning the present Example 3. 12 is a flowchart illustrating a process in which a recommendation unit according to the third embodiment creates search result data. It is a figure which shows the hardware constitutions of the computer which comprises the information search device concerning Examples 1-3. It is a figure for demonstrating the problem of a prior art.

Explanation of symbols

100, 200, 300 Information retrieval device 110, 210, 310 Metadata acquisition unit 120, 220, 320 Operation history recording unit 130, 230, 330 Preference information setting unit 140, 240, 340 Index creation unit 150, 250, 350 Recommendation unit 160, 260, 360 Exclusion history recording unit 170, 270, 370 Exclusion condition adjustment unit 180, 280, 380 Display unit 190, 290, 390 Search result recording unit 400 Computer 410 Input device 420 Output device 430 RAM
430a Various data 440 ROM
450 Media reader 460 Communication device 470 CPU
470a Search result creation process 480 HDD
480a Search result creation program 490 bus

Claims (5)

A metadata acquisition unit that acquires and stores metadata of information to be searched;
An operation history recording unit for storing operation history information in which the contents of the past operation history by the user and the metadata of the information to be operated are associated;
An index creation unit that creates each index from the metadata of the information to be searched stored in the metadata acquisition unit and the text part of the metadata of the operation target information stored in the operation history recording unit;
After obtaining each search result by comparing each index created by the index creation unit, the similarity between the index of the operation history information and the search result information, or the index of the search result information A recommendation unit that extracts a search result whose similarity is within a predetermined range and a comparison result other than the text portion of the metadata satisfies a predetermined condition as an excluded search result that is a candidate search result to be excluded;
An exclusion history recording unit for storing information of the exclusion search result extracted by the recommendation unit;
Based on the information on the exclusion search result stored in the exclusion history recording unit and the information on the operation history stored in the operation history recording unit, the condition for the recommendation unit to extract the exclusion search result is as follows. An exclusion condition adjustment section to be adjusted;
An information retrieval apparatus comprising:   The recommendation unit may include one of the compared search results when the similarity between the indexes of the search result information is within a predetermined range and a comparison result other than the text portion of the metadata satisfies a predetermined condition. The information search apparatus according to claim 1, wherein the search result is extracted as an excluded search result.   The exclusion condition adjustment unit compares the information of the exclusion search result with the information on the operation target stored in the operation history recording unit, and if the matching information is included, the similarity of the index is determined. The information search apparatus according to claim 1, wherein a predetermined range to be compared is narrowed.   The said exclusion condition adjustment part produces | generates the 1st exclusion information from which a similarity becomes more than predetermined value, and the 2nd exclusion information from which a similarity becomes less than a predetermined value, The said 2nd exclusion information and the said operation history recording part The predetermined range to be compared with the similarity of the index is expanded when the matching information is not included when the information on the operation target stored in the table is compared. Or the information search device of 3.   The recommendation unit compares the broadcast start time of the metadata between the operation history information and the search result information, or compares the broadcast start time of the metadata between the search results, and the search results having different broadcast start times The information search device according to any one of claims 1 to 4, wherein the information is extracted as the exclusion search result.

Images