JP2020030634A - Search device, search method, and search program - Google Patents

Abstract

To enhance probability to display search result data desired by a user in a high rank.SOLUTION: A plurality of search result data is extracted from a plurality of data to be searched. The plurality of search result data is outputted from an output apparatus. An attachment of a positive evaluation to selected search result data is received. Similarity information for identifying data to be searched that is similar to each data to be searched is generated. When an output order of the plurality of search result data is determined, the output order of second search result data similar to first search result data, that is identified based on the first search result data whose number of attached positive evaluations is equal to or greater than a first referential number and the similarity information, ranks higher than the output order of third search result data whose number of the attached positive evaluations is less than the first referential number and dissimilar to the first search result data.SELECTED DRAWING: Figure 3

Description

  The present invention relates to a search device, a search method, and a search program.

  The search device extracts search result data that matches a search condition from a plurality of search target data, and displays the extracted search result data on a display. In addition, when there is a plurality of search result data that match the search condition, the search device causes the display to display the plurality of search result data.

  When a plurality of search result data is displayed on the display, it is expected that the search result data desired by the user is displayed at a higher position.

  In the search engine sorting method described in Patent Literature 1, a user gives a score to all information and search results by showing opinions such as consent, likes, disagreement, and disagreement in a display list of search results. Do (summary). In addition, sorting is automatically performed according to the level of the score value, and the one with the highest score is displayed at the top (summary).

  In the document search system described in Patent Literature 2, a process for distinguishing between a document file that the user may view for business purposes and a document file that does not have the possibility based on the search history data of the user is performed. (Summary). In a subsequent user search process, a selection process is performed to exclude a document file having no possibility from the calculation target of the matching degree. In doing so, a choice is made after assessing the likelihood of being potentially ranked higher (summary).

JP 2006-505178 A JP 2012-53535 A

  The conventional search device has a problem that search result data desired by a user may not be displayed at a higher position.

  For example, in the search engine sorting method described in Patent Literature 1, search results for which a user does not show an opinion to raise an evaluation point are not displayed at the top. For this reason, in the search engine sorting method described in Patent Literature 1, the search result desired by the user may not be displayed at the top.

  Further, in the document search system described in Patent Literature 2, a document file specified based on the search history data of the user and not likely to be viewed by the user for business purposes is not displayed at the top. For this reason, in the document search system described in Patent Literature 2, the search result desired by the user may not be displayed at the top depending on the search history data of the user.

  The present invention has been made in view of these problems. An object of the present invention is to provide a search device, a search method, and a search program that increase the possibility that search result data desired by a user is displayed at a higher position.

  The present invention is directed to a search device.

  A plurality of search result data matching the search condition is extracted from the plurality of search target data. The plurality of search result data is output by the output device.

  Assignment of a positive evaluation to the search result data selected from the plurality of search result data is accepted.

  Similarity information that specifies search target data similar to each search target data included in the plurality of search target data is generated.

  The output order of the plurality of search result data when outputting the plurality of search result data to the output device is determined. The first search result data in which the number of assigned positive evaluations is equal to or greater than the first reference number, and the second search result data similar to the first search result data specified based on the similarity information The output order is set higher than the output order of the third search result data in which the number of assigned positive evaluations is smaller than the first reference number and is not similar to the first search result data.

  The present invention is also directed to a search method and a search program.

  According to the present invention, the first search result data to which a large number of positive evaluations are given and the second search result data similar to the first search result data are displayed at the top. Therefore, there is a high possibility that the search result data desired by the user is displayed at the top.

  The objects, features, aspects and advantages of the present invention will become more apparent from the following detailed description and the accompanying drawings.

FIG. 2 is a block diagram illustrating hardware of the search device according to the first embodiment. FIG. 2 is a block diagram illustrating a search device according to the first embodiment. FIG. 7 is a diagram illustrating an example of a change in display order of a plurality of search result data before and after feedback learning in the search device of the first embodiment. 5 is a flowchart illustrating a flow of acquisition of a plurality of search target data and learning of similar search target data performed in the search device of the first embodiment. 5 is a flowchart illustrating a flow of feedback learning performed in the search device of the first embodiment. FIG. 4 is a diagram illustrating an example of a table used for managing a search target document in the search device of the first embodiment. FIG. 4 is a diagram illustrating an example of a table used for managing a search target document in the search device of the first embodiment. FIG. 4 is a diagram illustrating an example of a table used for managing a search target document in the search device of the first embodiment. FIG. 4 is a diagram illustrating an example of a table used for managing similar documents in the search device of the first embodiment. FIG. 4 is a diagram illustrating an example of a table used for browsing similar documents in the search device of the first embodiment.

1. Hardware of Search Device FIG. 1 is a block diagram illustrating hardware of the search device according to the first embodiment.

  The search device 1000 illustrated in FIG. 1 is a personal computer (PC) on which a search program 1020 is installed. The search device 1000 includes a central processing unit (CPU) 1040, a memory 1042, a hard disk drive 1044, an input device 1046, and an output device 1048. The search device 1000 may include elements other than these elements.

  A search program 1020 is installed in the hard disk drive 1044. The search program 1020 is installed in the hard disk drive 1044 by writing data read from an external storage medium 1060 such as a compact disk (CD), a digital versatile disk (DVD), or a universal serial bus (USB) memory to the hard disk drive 1044. This may be performed, or may be performed by writing data received via the network 1080 to the hard disk drive 1044. The hard disk drive 1044 may be replaced with another type of auxiliary storage device. For example, the hard disk drive 1044 may be replaced with a solid state drive or a random access memory (RAM) disk. The hard disk drive 1044, the external storage medium 1060, the solid state drive, the RAM disk, and the like are computer-readable recording media on which the search program 1020 is recorded.

  The CPU 1040 loads the search program 1020 into the memory 1042, and executes the loaded search program 1020. Thus, the PC functions as the search device 1000.

  The input device 1046 is a keyboard, a mouse, a touchpad, a trackball, a touch panel, a switch, a dial, a microphone, or the like.

  The output device 1048 is a display, a printer, a speaker, or the like. In the following, description will be given on the assumption that the output device 1048 is a display, and the data is output by displaying the data on the display.

2 Search Device FIG. 2 is a block diagram illustrating a search device according to the first embodiment.

  As illustrated in FIG. 2, the search device 1000 includes an input device 1046, an output device 1048, an acquisition unit 1200, a storage unit 1202, a search unit 1204, an output control unit 1206, a reception unit 1208, a generation unit 1210, and a determination unit 1212. Is provided. The search device 1000 may include elements other than these elements.

  The acquisition unit 1200, the search unit 1204, the output control unit 1206, the reception unit 1208, the generation unit 1210, and the determination unit 1212 are configured by causing the CPU 1040 to execute a search program 1020. The storage unit 1202 includes at least one of the memory 1042 and the hard disk drive 1044.

  All or part of the processing performed by the CPU 1040 may be performed by another type of processing device. For example, all or a part of the processing performed by the CPU 1040 may be performed by a graphics processing device (GPU). All or a part of the processing performed by the CPU 1040 may be performed by hardware that does not execute a program.

  The acquisition unit 1200 acquires a plurality of search target data 1220 and stores the acquired plurality of search target data 1220 in the storage unit 1202.

  The search unit 1204 extracts search result data matching the search condition from the plurality of search target data 1220 stored in the storage unit 1202. In the following, description will be given assuming that there are a plurality of search result data 1240 that match the search conditions, and that the search unit 1204 extracts a plurality of search result data 1240 from the plurality of search target data 1220. The search unit 1204 may extract a plurality of search result data 1240 from a plurality of search target data stored in a storage unit provided in a computer connected to the search device 1000 via the network 1080.

  The output control unit 1206 causes the output device 1048 to output the extracted plurality of search result data 1240. The output control unit 1206 may output the plurality of search result data 1240 to an output device provided in a computer connected to the search device 1000 via the network 1080.

  In the following, the output device 1048 is a display, the output control unit 1206 is a display control unit, and outputting a plurality of search result data 1240 to the output device 1048 is performed by displaying a plurality of search result data 1240 on the display. The description proceeds assuming that the output order of the plurality of search result data 1240 is the display order of the plurality of search result data 1240.

  If the receiving unit 1208 detects that an operation of requesting to give “like” to the search result data selected from the displayed plurality of search result data 1240 has been performed on the input device 1046, Accepting “like” to the search result data is accepted. Also, when the receiving unit 1208 detects that an operation for requesting to give “bad” to the search result data selected from the plurality of displayed search result data 1240 has been performed on the input device 1046. Then, the application of “bad” to the search result data is accepted. The receiving unit 1208 also detects that an operation of requesting cancellation of the addition of “like” to the search result data selected from the plurality of displayed search result data 1240 has been performed on the input device 1046. In this case, cancellation of the addition of “like” to the search result data is accepted. If the receiving unit 1208 detects that an operation requesting the addition of “like” has been performed on an input device included in a computer connected to the search device 1000 via the network 1080, the reception unit 1208 may perform “like”. May be accepted. The same applies to the assignment of “bad” and the cancellation of the assignment of “like”. The receiving unit 1208 also generates feedback 1260 indicating the content of the assignment of “like”, the cancellation of the assignment of “like”, and the assignment of “bad”.

  The operation for requesting the addition of “like” is, for example, an operation of pressing a “like” button. The operation for requesting the “bad” is, for example, an operation of pressing a “bad” button. An operation for requesting cancellation of “like” is, for example, an operation of pressing a “like cancellation” button. Likes may be replaced by other types of positive ratings. For example, "like" may be replaced with "vote", "age", and the like. "Issue" may be replaced by another type of negative evaluation. For example, “bad” may be replaced with “opposite voting”, “sage”, and the like.

  The generation unit 1210 generates similarity information 1280 that specifies search target data similar to each search target data included in the plurality of search target data 1220 stored in the storage unit 1202. Search target data similar to each search target data is search target data other than each search target data included in the plurality of search target data 1220 stored in the storage unit 1202.

  The determination unit 1212 calculates display rank scores of the plurality of search result data 1240, and calculates a plurality of search result data 1240 on the display 1048 based on the calculated display rank scores of the plurality of search result data 1240. The display order 1290 of the search result data 1240 is determined.

  Further, when determining the display order 1290 of the plurality of search result data 1240, the determining unit 1212 learns the generated feedback 1260, and changes the display order 1290 of the plurality of search result data 1240 according to the learned feedback 1260. Change.

3. Change in Display Order of Plural Search Result Data Due to Feedback Learning FIG. 3 is a diagram illustrating an example of change in display order of a plurality of search result data before and after feedback learning in the search device of the first embodiment. .

  In the following, the numbers of “likes” and “bad” given to the search result data are referred to as “likes” and “bad” times of the search result data, respectively.

The search result data D ₁ , D ₂ , D ₃ , D ₄ , D ₅ ,..., D _n−1 , D _n illustrated in FIG. 3 are a plurality of search result data 1240 extracted by the search unit 1204. It is. The search result data D ₁ , D ₂ , D ₃ , D ₄ , D ₅ ,..., D _n−1 , D _n are search result data D ₁ , “Like” for which the number of “Likes” is three. "the number of times a is the search result data _{D 2} 1 times, and" good I "number of times is 0 times the search result data _{D 3, D 4, D 5} , ···, including the _{D n-1, D n.} Also, the search result data D ₁ , D ₂ , D ₃ , D ₄ , D ₅ ,..., D _n−1 , D _n are search result data D ₃ and D ₄ similar to the search result data D _1. Including. Also, the search result data D ₁ , D ₂ , D ₃ , D ₄ , D ₅ ,..., D _n−1 , D _n are search result data D _n in which the “bad” number of times is 1, and search result data _D 1 "bad I" count is 0 _{times, D 2, D 3, D} 4, D 5, ····, including the _{D n-1.}

As illustrated in FIG. 3, the determination unit 1212 determines the first search result data D ₁ and D ₂ whose number of “likes” is 1 or more, and the first search result data D ₁ similar to the _first search result data D ₁ . The display order of the second search result data D ₃ and D ₄ is changed to the third search result data D ₅ ,..., D that is less than one “like” times and is not similar to the first search result data D _{1. n-1,} to the upper than the display order of _{D n.} As a result, the first search result data D ₁ and D ₂ having a large number of “likes” and the second search result data D ₃ and D ₄ similar to the _first search result data D ₁ are associated with each other. Is displayed in the upper position with Therefore, there is a high possibility that the search result data desired by the user is displayed at the top, and the search can be executed easily and efficiently. Second search result data D ₃ and D _4, similar to the first search result data D ₁ is specified based on the generated similarity information 1280. For example, if the degree of similarity is the reference similarity above the second search result data D ₃ first search result data D ₁ of the first search result data D ₁ and the second search result data D ₃ It is determined to be similar, if the similarity is lower than the reference similarity is determined that the second search result data D ₃ not similar to the first search result data D _1. The first reference number to be compared with the number of “likes” may be other than one.

Further, determination unit 1212, the search result data D ₁ and D ₂ of the plurality of first is included if present, the plurality of first search result data D ₁ and D ₂ as shown in FIG. 3 As the number of “likes” of each first search result data increases, the display order of each first search result data becomes higher. Thus, the determination unit 1212, the first results of the display order data D ₁ the number of times "Like" is three times, the first number of searches "Like" is a one-time result display of data D ₂ Rank higher than ranking.

Further, as shown in FIG. 3, when there are a plurality of second search result data D ₃ and D ₄ , the determination unit 1212 determines whether the first search result data D ₁ and the plurality of second search results similarity between the second search result data is enough to each second top results display order of the data higher in the data D ₃ and D _4. Thus, the determination unit 1212, the second search results display order of the data D _3, the first search result relatively low data D ₁ having a first search result data D ₁ and a relatively high degree of similarity to higher order than display order of the second search result data D ₄ having a degree of similarity.

Further, determination unit 1212, as illustrated in Figure 3, the first search result display order of the data D ₁ and D _2, to higher than the display order of the second search result data D ₃ and D ₄ . First search result display order of the data D ₁ is in the higher order than display order of the second search result data D ₃ and D _4, the first search result display order of the data D ₂ is a second search result data D ₃ and may be lower than the display order of D _4. The relationship between the display order of the first search result data D ₁ and D _{2 and} the display order of the second search result data D ₃ and D ₄ can be adjusted by a display order score calculation algorithm described below.

Further, determination unit 1212, as illustrated in Figure 3, the third search result display order of the data D _n "bad ne" count is equal to or more than once, the first search result data D ₁ and D from the display order of ₂ and the second search result data D ₃ and D ₄ are lower. As a result, the "bad it" a large number of times the fourth search result data D _n is displayed in the lower. Therefore, there is a high possibility that unnecessary search result data is displayed at a lower position. The second reference number to be compared with the “bad” number may be other than one.

4. Acquisition of a plurality of search target data and learning of similar search target data FIG. 4 shows a flow of acquisition of a plurality of search target data and learning of similar search target data performed in the search device of the first embodiment. It is a flowchart illustrated.

  In step S101 illustrated in FIG. 4, the acquisition unit 1200 determines whether a trigger for acquiring all data has been input. The trigger for acquiring all data may be input according to a predetermined schedule or may be input manually. If it is determined that a trigger for acquiring all data has been input, steps S102, S103, and S104 are sequentially performed, and then step S105 is performed. On the other hand, when it is determined that the trigger for acquiring all data has not been input, step S105 is executed without executing steps S102, S103, and S104.

  In step S102, the acquisition unit 1200 starts crawling all data. The acquisition unit 1200 acquires all data during all data crawls, and causes the storage unit 1202 to store all acquired data. All the stored data are included in the plurality of search target data 1220. Further, the acquiring unit 1200 creates a search index of all acquired data.

  In step S103, the generation unit 1210 learns search target data similar to each search target data included in the plurality of search target data 1220 stored in the storage unit 1202, and searches for search target data similar to each search target data. Is generated. The generated similarity information 1280 includes a similarity score indicating the similarity between each search target data and search target data similar to each search target data.

  In step S104, the generation unit 1210 registers the similar information 1280 generated in step S103 in a database (DB) configured by the storage unit 1202.

  In step S105, the acquisition unit 1200 determines whether a trigger for acquiring difference data has been input. The trigger for acquiring the difference data may be input according to a predetermined schedule, or may be input manually. If it is determined that a trigger for acquiring difference data has been input, steps S106, S107, S108, and S109 are sequentially performed, and then step S110 is performed. On the other hand, if it is determined that the trigger for acquiring the difference data has not been input, step S110 is executed without executing steps S106, S107, S108 and S109.

  In step S106, the acquisition unit 1200 starts differential crawl. The acquisition unit 1200 acquires the difference data during the difference crawl. The difference data includes additional data, deleted data, and updated data.

  In step S107, the acquisition unit 1200 updates the plurality of search target data 1220 according to the acquired difference data. When updating the plurality of search target data 1220, the generation unit 1210 adds additional data to the plurality of search target data 1220, deletes deleted data from the plurality of search target data 1220, and updates the plurality of search target data The search target data included in the data 1220 is replaced.

  In step S108, the generation unit 1210 learns search target data similar to the updated search target data, and generates similarity information 1280 that specifies search target data similar to the updated search target data. The generated similarity information 1280 includes a similarity score indicating the similarity between the updated search target data and the search target data similar to the updated search target data. Also, the generation unit 1210 deletes the similarity information 1280 that specifies search target data similar to the deleted data.

  In step S109, the generation unit 1210 registers the similarity information 1280 generated in step S108 in the DB configured by the storage unit 1202.

  In step S110, it is determined whether or not search device 1000 has ended. If it is determined that the search device 1000 has not ended, step S101 is executed again. If it is determined that the search device 1000 has ended, the acquisition of a plurality of search target data and the learning of similar search target data end without executing step S101. Thus, steps S101 to S109 are repeatedly executed until the search device 1000 ends.

  According to steps S101 to S110, after the plurality of search target data 1220 is obtained, the plurality of search target data 1220 is updated. Therefore, the plurality of search target data 1220 can be kept up to date. Further, similarity information 1280 that specifies search target data similar to each search target data included in the plurality of search target data 1220 can be kept up to date.

5. Feedback Learning FIG. 5 is a flowchart illustrating a flow of feedback learning performed in the search device of the first embodiment.

  In step S121 illustrated in FIG. 5, the search unit 1204 performs a search on the plurality of search target data 1220 stored in the storage unit 1202, and matches the search condition from the plurality of search target data 1220. A plurality of search result data 1240 is extracted.

  In subsequent step S122, display control unit 1206 causes display 1048 to display the plurality of extracted search result data 1240.

  In subsequent step S123, receiving unit 1208 performs an operation on input device 1046 to request addition of “like” to the search result data selected from the plurality of search result data 1240 displayed on display 1048. It is determined whether or not it has been performed. Further, in step S124, reception unit 1208 determines whether or not an operation requesting cancellation of the addition of “like” to the search result data has been performed on input device 1046. In step S125, receiving unit 1208 determines whether or not an operation for requesting to give “bad” to the search result data is performed on input device 1046.

  If it is determined that an operation requesting the provision of “like” or an operation requesting cancellation of the provision of “like” is performed, steps S126, S127, and S128 are sequentially performed, and then steps S131 and S128 are performed. S132 is executed sequentially. When it is determined that the operation requesting the “bad” is performed, steps S129 and S130 are sequentially performed, and then steps S131 and S132 are sequentially performed. If it is determined that none of the operation requesting the grant of the "like", the operation requesting the cancellation of the grant of the "like", and the operation requesting the grant of the "bad" are performed, S132 is executed.

  In step S126, the determination unit 1212 updates the number of “likes” of the selected search result data. When it is determined that the operation requesting the addition of “like” is performed, the determination unit 1212 increases the number of “like” by one. On the other hand, when it is determined that the operation requesting cancellation of the assignment of “like” has been performed, the determination unit 1212 reduces the number of “like” by one.

  In step S127, the determination unit 1212 recalculates the display order score of the search result data similar to the selected search result data. The recalculated display ranking score includes “like count” of the selected search result data, and similarity indicating the similarity between the selected search result data and the search result data similar to the selected search result data. The degree score is reflected.

  In step S128, determination unit 1212 calculates a display rank score of the selected search result data. The calculated display rank score reflects the number of “likes” of the selected search result data and the display rank score of search result data similar to the selected search result data.

  In step S129, the determination unit 1212 updates the number of “bad” of the selected search result data. When it is determined that the operation requesting the “bad” is performed, the determining unit 1212 increases the number of “bad” by one.

  In step S130, determination unit 1212 calculates a display rank score of the selected search result data. The calculated display ranking score reflects the “number of bad times” of the selected search result data.

  In step S131, the determination unit 1212 determines the display order 1290 of the plurality of search result data 1240 based on the calculated display order score of the plurality of search result data 1240. Thus, when the next search is executed, the plurality of search result data 1240 is displayed on the display 1048 in accordance with the determined display order 1290 of the plurality of search result data 1240. The display order 1290 of the plurality of search result data 1240 displayed reflects the feedback 1260 from the user.

  In step S132, it is determined whether or not the user has logged out of search device 1000. If it is determined that the user has not logged out, step S121 is executed again. If it is determined that the user has logged out, the feedback learning ends without performing step S121. Thus, steps S121 to S131 are repeatedly executed until the user logs out of the search device 1000.

  According to steps S121 to S132, every time an operation requesting the grant of “like”, an operation requesting cancellation of the grant of “like”, or an operation requesting the grant of “bad” is performed, The display order 1290 of the plurality of search result data 1240 is updated.

6. Example of Table Used in Search Apparatus FIGS. 6, 7, and 8 are diagrams illustrating examples of tables used in managing search target documents in the search apparatus of the first embodiment.

  The tables illustrated in FIGS. 6, 7, and 8 are tables used when the plurality of search target data 1220 is a plurality of search target documents.

  The table 1300 illustrated in FIG. 6 includes a plurality of document identifiers (IDs) 1320, a plurality of document names 1322, a plurality of document uniform resource locators (URLs) 1324, a plurality of “likes” counts 1326, and a plurality of “likes”. The bad number 1328 is described. The plurality of document names 1322 are respectively associated with the plurality of document IDs 1320. A plurality of document URLs 1324 are associated with a plurality of document IDs 1320, respectively. The plurality of “like” times 1326 are respectively associated with the plurality of document IDs 1320. A plurality of “bad” times 1328 are respectively associated with a plurality of document IDs 1320.

  The plurality of document IDs 1320 identify a plurality of search target documents, respectively. Each document name included in the plurality of document names 1322 is the name of a document identified by a document ID associated with each document name. Each document URL included in the plurality of document URLs 1324 is a URL of a document identified by a document ID associated with each document URL. The number of “likes” included in the plurality of “likes” times 1326 is the number of “likes” of the document identified by the document ID associated with each “like” number. Each “poor” count included in the plurality of “poor” counts 1328 is the “poor” count of the document identified by the document ID associated with each “poor” count. The plurality of “likes” times 1326 and the plurality of “bad” times 1328 are used in learning the feedback 1260.

  The plurality of “like” times 1326 and the plurality of “bad” times 1328 are shared by a plurality of users. As a result, the information to be obtained is shared by a plurality of users, and a user without sufficient knowledge can obtain the information to be obtained.

  In a table 1400 shown in FIG. 7, a plurality of document file paths 1420 are described. The plurality of document file paths 1420 are paths of files in which a plurality of search target documents are described.

  In a table 1500 illustrated in FIG. 8, paths 1520 of a plurality of morphological analysis result files are described. The paths 1520 of the plurality of morphological analysis result files are paths of files describing the morphological analysis results of the plurality of search target documents, respectively. Morphological analysis results of a plurality of search target documents are used for learning similar search target data.

  FIG. 9 is a diagram illustrating an example of a table used for managing similar documents in the search device of the first embodiment.

  The table 1600 illustrated in FIG. 9 is a table used when the plurality of search target data 1220 is a plurality of search target documents.

  The table 1600 illustrated in FIG. 9 describes a plurality of document IDs 1620, a plurality of similar document IDs 1622, and a plurality of similarities 1624. The plurality of similar document IDs 1620 are respectively associated with the plurality of document IDs 1620. The plurality of similarities 1624 are respectively associated with the plurality of document IDs 1620.

  Each similar document ID included in the plurality of similar document IDs 1620 is a document ID for identifying a document similar to the document identified by the document ID associated with each similar document ID. Each similarity included in the plurality of similarities 1624 is identified by a document identified by the document ID associated with each similarity and a similar document ID associated with the document ID associated with each similarity. The degree of similarity with the document to be processed.

  FIG. 10 illustrates an example of a table used for browsing similar documents in the search device of the first embodiment.

  The table 1700 illustrated in FIG. 10 is a table used when the plurality of search target data 1220 is a plurality of search target documents.

  A table 1700 illustrated in FIG. 10 describes a plurality of document IDs 1720, a plurality of similar document IDs 1722, a plurality of document URLs 1724, a plurality of similar document URLs 1726, and a plurality of similarities 1728. The plurality of document IDs 1720, the plurality of similar document IDs 1722, and the plurality of similarities 1728 are respectively similar to the plurality of document IDs 1620, the plurality of similar document IDs 1622, and the plurality of similarities 1624 shown in FIG. The plurality of document URLs 1724 are respectively associated with the plurality of document IDs 1720. The plurality of similar document URLs 1726 are respectively associated with the plurality of similar document IDs 1720. Each document URL included in the plurality of document URLs 1724 is a URL of a document identified by a document ID associated with each document URL. Each similar document URL included in the plurality of similar document URLs 1726 is a URL of a document identified by a similar document ID associated with each similar document URL.

  Although the present invention has been described in detail, the above description is illustrative in all aspects and the present invention is not limited thereto. It is understood that innumerable modifications that are not illustrated can be assumed without departing from the scope of the present invention.

1000 Search device 1046 Input device 1048 Output device (display)
1200 acquisition unit 1202 storage unit 1204 search unit 1206 output control unit (display control unit)
1208 reception unit 1210 generating unit 1212 determination unit 1220 multiple retrieval target data 1240 multiple retrieval result data 1260 feedback 1280 similar information 1290 display order _{D 1, D 2, D 3} , D 4, D 5, ···, D n _{-1, D n} search result data

Claims (10)

A search unit that extracts a plurality of search result data that match a search condition from a plurality of search target data;
An output control unit that outputs the plurality of search result data to an output device,
An accepting unit that accepts the assignment of a positive evaluation to the search result data selected from the plurality of search result data,
A generation unit that generates similar information that specifies search target data similar to each search target data included in the plurality of search target data,
A first search result in which the output order of the plurality of search result data when outputting the plurality of search result data to the output device is determined, and the number of positive evaluations given is equal to or greater than a first reference number. The output rank of the second search result data similar to the first search result data, which is specified based on the data and the similarity information, is set such that the number of positive evaluations given is greater than the first reference number. A determining unit that ranks higher than the output order of the third search result data that is at least not similar to the first search result data;
A search device comprising: 2. The search device according to claim 1, wherein the accepting unit accepts cancellation of assignment of a positive evaluation to search result data selected from the plurality of search result data. 3. The receiving unit receives a negative evaluation given to search result data selected from the plurality of search result data,
The determining unit may determine the output order of the fourth search result data in which the number of assigned negative evaluations is equal to or greater than a second reference number, by outputting the first search result data and the second search result data. 3. The search device according to claim 1, wherein the search device is lower than the order. The determining unit, when there are a plurality of first search result data, sets an output order of each of the first search result data as the number of positive evaluations given to each first search result data increases. 4. The search device according to claim 1, wherein the search device is a higher rank device. The determining unit, when there are a plurality of second search result data, the higher the similarity between the first search result data and each second search result data, the higher the similarity between each second search result data. The search device according to any one of claims 1 to 4, wherein the output order is set higher. The determining unit includes:
Calculating a display ranking score of the plurality of search result data;
The number of positive evaluations given to the search result data included in the plurality of search result data, and the similarity between the search result data and the search result data similar to the search result data are similar to the search result data. Of the search result data to be displayed,
The number of positive evaluations given to the search result data, and the display rank score of search result data similar to the search result data, reflected in the display rank score of the search result data,
The search device according to any one of claims 1 to 5, wherein an output order of the plurality of search result data is determined based on a display order score of the plurality of search result data. The receiving unit receives a negative evaluation given to search result data selected from the plurality of search result data,
The search device according to claim 6, wherein the determining unit reflects the number of negative evaluations given to the search result data in a display rank score of the search result data. 8. The search device according to claim 1, wherein the determination unit sets the output order of the first search result data higher than the output order of the second search result data. 9. a) extracting a plurality of search result data that match a search condition from a plurality of search target data;
b) outputting the plurality of search result data to an output device;
c) receiving a positive evaluation given to the search result data selected from the plurality of search result data;
d) generating similarity information identifying search target data similar to each search target data included in the plurality of search target data,
e) determining the output order of the plurality of search result data when outputting the plurality of search result data to the output device, wherein the number of assigned positive evaluations is equal to or greater than a first reference number; The output order of the second search result data similar to the first search result data, which is specified based on the search result data and the similarity information, is determined based on the number of positive evaluations assigned to the first reference. Raising the output order of third search result data that is less than the number and is not similar to the first search result data;
A search method comprising: a) extracting a plurality of search result data that match a search condition from a plurality of search target data;
b) outputting the plurality of search result data to an output device;
c) receiving a positive evaluation given to the search result data selected from the plurality of search result data;
d) generating similarity information identifying search target data similar to each search target data included in the plurality of search target data,
e) determining the output order of the plurality of search result data when outputting the plurality of search result data to the output device, wherein the number of assigned positive evaluations is equal to or greater than a first reference number; The output order of the second search result data similar to the first search result data, which is specified based on the search result data and the similarity information, is determined based on the number of positive evaluations assigned to the first reference. Raising the output order of third search result data that is less than the number and is not similar to the first search result data;
A search program that causes a computer to execute.

Images