JP2013206066A - Data retrieval system and data retrieval method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate a retrieval result whose retrieval order is appropriately evaluated even when a retrieval result related to metadata and a retrieval result related to retrieval object data are separately acquired.SOLUTION: A data retrieval system 1 comprises: a retrieval result acquisition part 33 for acquiring a data retrieval result indicating extraction data from a plurality of retrieval object data; a metadata storage part 38a for storing a plurality of metadata; a data retrieval part 32 for generating a metadata retrieval result indicating extraction metadata from a plurality of metadata; a relation degree calculation part 34 for calculating an extraction data relation degree and an extraction metadata relation degree; a retrieval result generation part 36 for calculating an integrated relation degree in which the extraction data relation degree and the extraction metadata relation degree are incorporated into the plurality of retrieval object data; and a data communication part 37 for outputting an integrated retrieval result in which the data retrieval result and the metadata retrieval result are integrated, in order based on the integrated relation degree.

Description

  The present invention relates to a data search system and a data search method for searching for data via a network.

  Conventionally, a service for searching data in a communication network such as the Internet using a search keyword has been provided. With regard to the technology related to such services, for example, the following Patent Document 1 discloses a distributed type in which search results obtained by a plurality of search execution devices are integrated and output by weighting according to the quality of information. A search method is disclosed. According to such a function, a result suitable for the search request can be obtained as a search result in the distributed search device.

JP 2003-248691 A

  However, in the conventional search method described above, the storage location of the data itself is different from the storage location of the additional data such as metadata added to the data, and the search results regarding the data itself and the additional data are obtained separately. In addition, it is difficult to obtain a search result that appropriately evaluates the relevance of each search result. For example, when metadata matches the search keyword for some data and both the metadata and the data itself match the search keyword for other data, the search ranking of both data tends not to be evaluated properly It is in.

  Therefore, the present invention has been made in view of such problems, and even when a search result related to additional data added to search target data and a search result related to search target data are obtained separately, the present invention is appropriate. It is an object of the present invention to provide a data search system and a data search method capable of generating a search result whose search rank is evaluated.

  In order to solve the above problems, a data search system according to the present invention includes a search result acquisition unit that acquires a first search result indicating a plurality of extracted data extracted by a search keyword from a plurality of search target data, and a plurality of search result acquisition means. Additional data storage means for storing a plurality of additional data added to the search target data, and extracting a plurality of extracted additional data related to the search keyword from the plurality of additional data, and adding the plurality of extracted additional data A data search unit for generating a second search result indicating the plurality of search target data, a search keyword for the plurality of extracted data based on the first search result acquired by the first search result acquisition unit, and A plurality of search target data corresponding to the plurality of extracted additional data based on the second search result acquired by the data search means. Relevance calculation means for calculating a second relevance level with the search keyword, and a plurality of extracted data and a plurality of search target data corresponding to the plurality of extracted additional data. The search result generation means for calculating the third relevance level taking into account the relevance level of the first search result and the third search result obtained by combining the first search result and the second search result are based on the third relevance level. Output means for outputting in order.

  Alternatively, in the data search method of the present invention, the search result acquisition means acquires a first search result indicating a plurality of extracted data extracted by a search keyword from a plurality of search target data; An additional data storage step in which the additional data storage means stores a plurality of additional data added to the plurality of search target data; and a plurality of extracted additional data related to the search keyword from the plurality of additional data. And a data search step for generating a second search result indicating a plurality of search target data to which a plurality of extraction additional data is added, and a relevance calculation means are acquired by the first search result acquisition means. Based on the first search result, the first degree of association with the search keyword of the plurality of extracted data is calculated and acquired by the data search means A relevance level calculating step for calculating a second relevance level with a search keyword of a plurality of search target data corresponding to a plurality of extracted additional data based on the search result of 2; A search result generating step for calculating a third relevance factor including the first relevance factor and the second relevance factor for the data and a plurality of search object data corresponding to the plurality of additional data to be extracted; Output a third search result obtained by combining the first search result and the second search result in a rank based on the third relevance level.

  According to such a data search system or data search method, a first search result searched for a plurality of search target data based on a search keyword is acquired, and a plurality of search results are acquired based on the same search keyword. A second search result searched for the additional data is generated. Then, the first relevance of the plurality of extracted data is calculated based on the first search result, and the second relevance of the plurality of search target data corresponding to the plurality of extracted additional data is calculated based on the second search result. After the degree is generated, a third relevance level in which the first relevance level and the second relevance level are added is calculated for each of a plurality of search target data included in both search results. The combined third search result is output in the order based on the third degree of association. Thus, even when a search for a plurality of search target data and a search result for additional data corresponding to the search target data are obtained separately, the search keyword in both search results By adding the relevance level, a search result in which the relevance level for each search target data is appropriately evaluated can be obtained. As a result, it is possible to generate a search result in which the search rank is appropriately evaluated for the search target data.

  The relevance calculation means preferably calculates the first relevance based on the number of appearances of the search keyword in the plurality of extracted data. In this way, the relevance of the extracted data in the first search result can be easily obtained.

  It is also preferable that the relevance degree calculating means calculates the first relevance degree based on the ranks of the plurality of extracted data in the first search result. Even in this way, the relevance of the extracted data in the first search result can be easily obtained.

  Further, the search result generating means weights and adds the first relevance level and the second relevance level to the plurality of search target data corresponding to the plurality of extracted data and the plurality of additional data to be extracted. It is also preferable to calculate relevance. In this case, the degree of association for each search target data can be more appropriately evaluated using the degree of association with the search keyword in the first and second search results.

  It is also preferable to further include parameter changing means for dynamically changing a weighting value used at the time of weighting addition based on a user data selection history for the third search result output by the output means. With such parameter changing means, it is possible to generate a search result having a high utility value for the user.

  Furthermore, the parameter changing means changes the weighting value based on a comparison between the number of times of selecting the first selection result indicated by the data selection history and the number of times of selecting the second selection result indicated by the data selection history. It is also preferable. In this way, a search result having a high utility value for the user can be generated more reliably.

  According to the present invention, even when a search result related to additional data added to search target data and a search result related to search target data are obtained separately, a search result whose search rank is appropriately evaluated is generated. be able to.

1 is a schematic configuration diagram of a data search system 1 according to a preferred embodiment of the present invention. It is a block diagram which shows the hardware constitutions of the information processing apparatus which comprises the terminal device 2 and the metadata search server apparatus 3 of FIG. It is a figure which shows an example of the data structure of the metadata stored in the metadata storage part 38a of FIG. It is a figure which shows an example of a data structure of the integrated search result data stored in the selection history storage part 38c of FIG. It is a figure which shows an example of a data structure of the search object data stored in the data storage part 43a of FIG. It is a flowchart which shows the operation | movement at the time of the integrated search result data generation by the data search system 1 of FIG. It is a figure which shows the example of an output of the integrated search result by the data search system 1 of FIG. It is a schematic block diagram of the data search system 101 concerning the modification of this invention.

  Hereinafter, preferred embodiments of a data search system and a data search method according to the present invention will be described in detail with reference to the drawings. In the description of the drawings, the same elements are denoted by the same reference numerals, and redundant description is omitted.

  FIG. 1 is a schematic configuration diagram of a data search system 1 according to a preferred embodiment of the present invention. A data search system 1 shown in FIG. 1 manages search target data such as data stored by a user using a terminal device 2 and data published on the Internet, and executes search processing of search target data. System. Examples of such search target data include schedule data, photo data, document data, SNS (Social Networking Service) posting data, e-mail data, music data, news data, and the like. More specifically, the data search system 1 includes a terminal device 2 used by a user who intends to execute data search, a metadata search server device 3 that manages metadata (additional data) added to search target data, and the like. It is comprised by. The terminal device 2 and the metadata search server device 3 can perform data communication with each other via a communication network NW configured by a mobile communication network, a wired communication network, or the like adopting a mobile communication method. It is possible. Further, the terminal device 2 and the metadata search server device 3 can perform data communication with the data search server device 4 that manages the search target data via the communication network NW. The metadata search server device 3 and the data search server device 4 may be configured by a single server device or a server system in which a plurality of server devices operate in cooperation. The terminal device 2 is a terminal device represented by a mobile phone terminal, a smartphone, a PDA, or the like.

  FIG. 2 is a block diagram showing a hardware configuration of an information processing apparatus constituting the terminal device 2 or the metadata search server device 3 of the data search system 1 of FIG. The information processing apparatus 100 physically includes a CPU 51, a RAM 52 and a ROM 53 that are main storage devices, an auxiliary storage device 56 such as a hard disk device, and input devices 55 such as input keys, a touch panel, and a mouse that are input devices. And an output device 57 such as a display and a speaker, and a communication module 54 that controls transmission / reception of data to / from other terminal devices and server devices via the communication network NW. The functions realized by the terminal device 2 or the metadata search server device 3 are such that a predetermined program is read on the hardware such as the CPU 51 and the RAM 52 shown in FIG. This is realized by operating the input device 55 and the output device 57 and reading and writing data in the RAM 52 and the auxiliary storage device 56.

  Returning to FIG. 1, the metadata search server device 3 includes, as functional components, an index creation unit 31, a data search unit (data search unit) 32, a search result acquisition unit (search result acquisition unit) 33, , Relevance calculation unit (relevance calculation unit) 34, parameter change unit (parameter change unit) 35, search result generation unit (search result generation unit) 36, data communication unit (output unit) 37, metadata A storage unit 38a, an index storage unit 38b, and a selection history storage unit 38c are provided.

  First, the function of each component of the metadata search server device 3 will be described in detail.

  The index creation unit 31 creates an index for searching the metadata stored in the metadata storage unit 38a when a request from the user of the terminal device 2 is received. This index is referred to in order to speed up the search process when searching for metadata stored in the metadata storage unit 38a by the data search unit 32.

  When the data search unit 32 receives a search query (search request) from the terminal device 2, the data search unit 32 selects the search query from among the metadata added to the plurality of search target data stored in the metadata storage unit 38a. A plurality of metadata related to the included search keyword is extracted as extracted metadata. At this time, the data search unit 32 executes a data search process while referring to the index stored in the index storage unit 38b.

  Here, FIG. 3 shows an example of the configuration of the metadata stored in the metadata storage unit 38a. As shown in the figure, a plurality of metadata is stored in the metadata storage unit 38a, and these metadata are each one of a plurality of search target data stored in the data search server device 4. Are generated and stored in advance corresponding to each. For example, as metadata having a plurality of items, “creation date / time: 2012/1/30 11:00”, “creation position: Kita-ku, Okayama city, Okayama”, “schedule at creation: Okayama business trip A company meeting”, “ Accompanying person: Suzuki, “File name: Meeting material 1”, “Keyword: SNS service survey” and other data items are stored in association with ID “d001” for identifying the corresponding search target data. Has been. As described above, the metadata storage unit 38 a stores a plurality of metadata added to a plurality of search target data stored in the data search server device 4. Such metadata is automatically assigned to search target data by the terminal device 2 such as “creation date”, “creation position”, “schedule at creation”, “accompanying person”, and “file name”. And two types of items registered by the user in the terminal device 2 such as “keyword”. The data search unit 32 extracts IDs including a search keyword in the data of any item of the metadata as a plurality of IDs for specifying the extracted metadata while referring to the metadata as shown in FIG. To do. For example, when the search keyword is “SNS”, two IDs “d001” and “d002” including “SNS” in the item “keyword” are extracted from the metadata as shown in FIG. Then, the data search unit 32 creates a metadata search result indicating a plurality of IDs specifying the extracted metadata, and delivers it to the relevance calculation unit 34. The data search unit 32 creates the metadata search result by arranging IDs according to the search order determined based on the degree of coincidence with the search keyword or the number of occurrences of the search keyword.

  Here, when the date and time information is included in the search keyword, the data search unit 32 extracts metadata having items that are temporally approximated to the time indicated by the date and time information. For example, if date and time information “2012/1/29” is included in the search keyword, the range of “2012/1/28” to “2012/1/30” is one day before and after the date indicated by the date and time information. Extract metadata including time information. In addition, when the search keyword includes position information, the data search unit 32 extracts metadata having items that spatially approximate the position indicated by the position information. For example, if the search keyword includes the location information “Okayama City, Okayama Prefecture”, the location name may be matched (including partly matched), or the location name may be converted into latitude and longitude. Judgment of close proximity (whether or not the position is approximate) may be performed.

  Returning to FIG. 1, when the search result acquisition unit 33 receives a search query from the terminal device 2, the search result acquisition unit 33 transfers the search query to the data search server device 4. Acquire data search results regarding the extracted data extracted from the inside. This data search result is created by arranging IDs according to a search order determined by a predetermined method in the data search server device 4. Here, various existing algorithms can be used as a method for determining the search order. The data search result of the data search server device 4 may include the number of occurrences of the search keyword in each extracted data. For example, if the search order is the order of ID “d001”, “d002”, and “d004”, the IDs arranged in this order together with the number of occurrences of the search keyword in each extracted data are the data search results. Included. Then, the search result acquisition unit 33 delivers the acquired data search result to the relevance calculation unit 34.

The degree-of-association calculation unit 34 calculates the degree of association (extraction data relevance) for the search keyword related to each piece of extracted data extracted by the data search server device 4 based on the data search result acquired by the search result acquisition unit 33. To do. Further, based on the metadata search result created by the data search unit 32, the degree of relevance (extraction metadata relevance) of the search target data corresponding to each extracted metadata extracted by the data search unit 32 to the search keyword is calculated. calculate. In other words, the degree-of-association calculation unit 34, based on the number of occurrences N _K of the search keyword in the respective extraction data, the total number N _KTOTAL number of occurrences of search terms in all of the search target data, the following formula;
V _R1 = N _K / N _{KTOTAL
Is used} to calculate the relevance VR1 of each extracted data. Moreover, the degree-of-association calculation unit 34 calculates the relevance V _R2 of the extracted metadata and all the fixed value 1.

Here, the relevance calculation unit 34 uses the following formula based on the number N _{D of} extracted data included in the data search result and the search order N _ORDER of the corresponding extracted data instead of the above formula:
_{V R1 = (N D -N ORDER} ) / N D
With, it may calculate the relevance V _R1 of the extraction data. Further, the relevance level calculating unit 34 may calculate the extracted metadata relevance level _{VR2 in} the same manner as the extraction data relevance level _VR1 . The degree-of-association calculation unit 34 passes the calculated extracted data relevance level _VR1 and extracted metadata relevance level _VR2 to the search result generation unit 36 together with an ID for identifying the corresponding search target data.

The search result generation unit 36 extracts the extracted data relevance level _VR1 and the extraction for each of a plurality of search target data corresponding to a plurality of extracted data included in the data search result and a plurality of extracted metadata included in the metadata search result. calculating the integrated relevance V _R3 in consideration and metadata relevance V _R2. That is, the search result generation unit 36 extracts the relevance level corresponding to the ID of the same search target data from the acquired extracted data relevance level V _R1 and extracted metadata relevance level V _R2 , and both relevance levels V _R1 , V _R2 is integrated by weighted addition to determine the integrated relevance _VR3 . Specifically, for each ID included in the data search result and the metadata search result, using the variable coefficient α (<1) for weighting addition, the following formula:
V _R3 = V _R1 × α + V _R2 × (1-α)
Accordingly, calculating the integrated relevance _{V R3.} At this time, the search result generating unit 36 uses the variable coefficient α set by the parameter changing unit 35. For search target data included in the data search result but not included in the metadata search result, V _R2 = 0 is set, and the search target data included in the metadata search result and not included in the data search result is included. On the other hand, the integrated relevance V _R3 is obtained with V _R1 = 0. Then, the search result generation unit 36 stores the integrated relevance _VR3 calculated for each ID of all search target data included in the data search result and the metadata search result in the selection history storage unit 38c.

  FIG. 4 shows an example of the data structure of the integration relevance stored in the selection history storage unit 38c by the search result generation unit 36. As shown in the figure, for each ID “d001”... Included in the data search result and metadata search result, the extracted data relevance “0.8”... And the extracted metadata relevance “ 1 ”... And the integrated relevance“ 0.8 * 0.3 + 1 * 0.7 = 0.94 ”(when α = 0.3) calculated by the search result generation unit 36 are stored in association with each other. The data stored in the selection history storage unit 38c is used as integrated search result data to be presented to the terminal device 2 according to the search query.

  Returning to FIG. 1, the data communication unit 37 executes data communication between the metadata search server device 3 and the terminal device 2 and the data search server device 4. In particular, the data communication unit 37 receives the search query from the terminal device 2, reads out the integrated search result data that combines the data search result and the metadata search result from the selection history storage unit 38c, and transmits the search query. A reply is sent to the terminal device 2. Further, the data communication unit 37 monitors whether or not a reference request related to the search target data presented by the integrated search result data is received from the terminal device 2, and the number of reference requests (selection history) is determined for each search target data. Stored in the selection history storage unit 38c. For example, as shown in FIG. 4, the data communication unit 37 sets a selection number “N1”... Indicating the number of reference requests by the user of the terminal device 2 for each of a plurality of IDs “d001”. Record. The data communication unit 37 accumulates and stores the number of selections in a predetermined period as history data regarding the integrated search result data created for a plurality of search queries. Further, the data communication unit 37 transfers the search query to the data search server device 4 and acquires the data search result from the data search server device 4 accordingly.

Parameter changing unit 35, the search result generation unit 36 the coefficient α to be referred to for calculating the integrated relevance V _R3, changed according to the data selection history by the user of the search target data stored in the selection history storage unit 38c To do. That is, the parameter changing unit 35 refers to the number of selections in a predetermined period included in the selection history storage unit 38c, and totals the total number of selections N _S1 that selected the search target data included in the metadata search result. The total _NS2 of the number of times of selecting the search target data included in the data search result is totaled. For example, according to the example of FIG. 4, the number of selections “N1”, “N2”, “N4” associated with an ID having extracted data relevance greater than 0 (ie, an ID included in the data search result) In total, the total value “N _S1 = N1 + N2 + N4” is calculated, and the number of selections “N1” associated with the ID having the extracted metadata relevance degree exceeding 0 (that is, the ID included in the metadata search result), “N2” and “N3” are added together to calculate a total value “N _S2 = N1 + N2 + N3”. Then, the parameter changing unit 35 sets the coefficient α to the following formula:
α = N _S1 / (N _S1 + N _S2 )
Is used to dynamically change the two total values N _S1 and N _S2 of the number of selections.

  Next, the functional configuration of the terminal device 2 will be described. The terminal device 2 includes a metadata registration unit 21, a query input unit 22, an information output unit 23, and a data communication unit 24. The metadata registration unit 21 receives registration of some items of metadata stored in the metadata storage unit 38 a from the user, and transmits data related to the registration to the metadata search server device 3 via the data communication unit 24. . The query input unit 22 receives an input of a search query from the user and transmits it to the metadata search server device 3 via the data communication unit 24. The information output unit 23 receives the integrated search result data from the metadata search server device 3 and outputs it to an output device such as a display. The data communication unit 24 performs data communication with the metadata search server device 3 and the data search server device 4. In addition, when the data communication unit 24 receives a selection input for data reference with respect to the search target data presented by the integrated search result data from the user, the data communication unit 24 sends the selection input to the metadata search server device 3 and the data search server. It transmits to the server device 4.

  Further, the functional configuration of the data search server device 4 will be described. The data search server device 4 includes a data search unit 41, a data communication unit 42, a data storage unit 43a, and an index storage unit 43b. As shown in FIG. 5, the data storage unit 43a stores a plurality of search target data “meeting material 1.doc”... In association with ID “d001”. The storage unit 43b stores a search index for search target data stored in the data storage unit 43a. The data search unit 41 extracts the extracted data from the search target data stored in the data storage unit 43a when the search query is received from the metadata search server device 3, and performs a data search indicating the extracted data. Generate results. Here, the data search unit 41 generates the data search result including the number of occurrences of the search keyword in each extracted data so that IDs indicating a plurality of extracted data are arranged in accordance with the search order. The data communication unit 42 performs data communication between the terminal device 2 and the metadata search server device 3.

  Hereinafter, the operation of the data search system 1 will be described with reference to FIG. 6, and the data search method in the data search system 1 will be described in detail. FIG. 2 is a flowchart showing an operation when the integrated search result data is generated by the data search system 1.

  First, the terminal device 2 accepts a search query for requesting a search for search target data using a search keyword (step S101). Then, the search query is accepted as a metadata search request by the data search unit 32 of the metadata search server device 3 (step S102). The data search unit 32 extracts a plurality of metadata stored in the metadata storage unit 38a using the search keyword, thereby acquiring a metadata search result (step S103). Further, the search result acquisition unit 33 of the metadata search server device 3 transmits a search query as a data search request to the data search server device 4 (step S104). In response to this, the search result acquisition unit 33 acquires the data search result related to the search target data using the search keyword from the data search server device 4 (step S105).

  Thereafter, the relevance calculation unit 34 of the metadata search server device 3 determines the extracted data relevance regarding each extracted data and the extracted metadata relevance regarding each extracted metadata based on the data search result and the metadata search result. Calculated (step S106). Next, the parameter change unit 35 of the metadata search server device 3 refers to the selection history of the search target data stored in the selection history storage unit 38c, so that the coefficient α used in the calculation of the degree of integration is calculated. The value is dynamically changed (step S107). Further, the integrated result is calculated by weighting and adding the extracted data relevance and the extracted metadata relevance for each search target data by the search result generation unit 36 of the metadata search server device 3 (step S108). Thereafter, the search result generation unit 36 stores the calculated integrated relevance in the selection history storage unit 38c as integrated search result data associated with an ID for specifying the search target data (step S109). Finally, the integrated search result data stored in the selection history storage unit 38c is transmitted to the terminal device 2 by the data communication unit 37 of the metadata search server device 3 (step S110).

FIG. 7 shows an example of the output screen D ₁ displayed on the terminal device 2 in accordance with the integrated search result data transmitted from the metadata search server device 3. As shown in the figure, the search target data “meeting material 1.doc”, “meeting material 2.doc”, and “meeting material 4.doc” are extracted as integrated search results for the search keyword “SNS”. Indicated. Also, the integrated search result includes a related part “SNS service survey” in the metadata hit with the search keyword and a related part “Search the SNS market size in 2012…” in the search target data hit the search keyword. These are displayed together for each search target data.

  According to the data search system 1 and the data search method in the data search system 1 described above, the data search result searched for a plurality of search target data based on the search keyword is acquired, and the same search keyword is obtained. Based on the plurality of metadata, a metadata search result is generated. Then, the extracted data relevance is calculated based on the data search result, and the extracted metadata relevance is generated based on the metadata search result. Thereafter, the integrated relevance is calculated for each of a plurality of search target data included in both search results, and the integrated search results obtained by combining both search results are output in the order based on the integrated relevance. As a result, even if the search for multiple search target data and the search results for the metadata corresponding to the search target data are obtained on separate devices, the search in both search results By taking into account the degree of association with the keyword, a search result in which the degree of association with the search keyword for each search target data is appropriately evaluated can be obtained. As a result, it is possible to generate a search result in which the search order is appropriately evaluated for the relevance of the search keyword to the search target data.

  Moreover, since the extracted data relevance is calculated based on the number of appearances of the search keyword in the extracted data and the search rank in the data search result, the extracted data relevance in the data search result can be easily obtained.

  Furthermore, since the integrated relevance is obtained by weighted addition of the extracted data relevance and the extracted metadata relevance, the relevance with the search keyword in the data search result and the metadata search result is used for each search target data. Relevance can be evaluated more appropriately. Further, since the coefficient α used for the weighted addition is dynamically changed based on the user's data selection history for the integrated search result, a search result having high utility value for the user can be generated.

  In addition, this invention is not limited to embodiment mentioned above.

  For example, some or all of the components of the metadata search server device 3 as shown in FIG. 1 may be included in the terminal device used by the user. For example, as illustrated in FIG. 8, a search result acquisition unit 33, a relevance calculation unit 34, a parameter change unit 35, a search result generation unit 36, and a selection history storage unit 38 c may search for search target data. You may be provided in the terminal device 102 to be used.

  DESCRIPTION OF SYMBOLS 1,101 ... Data search system, 2,102 ... Terminal device, 3,103 ... Server device for metadata search, 32 ... Data search part (data search means), 33 ... Search result acquisition part (search result acquisition means), 34... Relevance calculation unit (relevance calculation unit) 35. Parameter change unit (parameter change unit) 36. Search result generation unit (search result generation unit) 38 a. Metadata storage unit (additional data storage unit).

Claims (7)

Search result acquisition means for acquiring a first search result indicating a plurality of extracted data extracted from a plurality of search target data by a search keyword;
Additional data storage means for storing a plurality of additional data added to the plurality of search target data;
Data for extracting a plurality of extracted additional data related to the search keyword from the plurality of additional data and generating a second search result indicating a plurality of search target data to which the plurality of extracted additional data is added Search means;
Based on the first search result acquired by the first search result acquisition means, a first relevance level of the plurality of extracted data with the search keyword is calculated, and the first search result acquired by the data search means is obtained. Relevance calculation means for calculating a second relevance of the plurality of search target data corresponding to the plurality of extracted additional data with the search keyword based on the search result of 2;
A third degree of association is calculated by adding the first degree of association and the second degree of association to the plurality of extracted data and the plurality of search target data corresponding to the plurality of extracted additional data. Search result generation means;
Output means for outputting a third search result obtained by combining the first search result and the second search result in an order based on the third degree of association;
A data retrieval system comprising: The relevance calculation means includes:
Calculating the first relevance based on the number of appearances of the search keyword in the plurality of extracted data;
The data search system according to claim 1. The relevance calculation means includes:
Calculating the first relevance based on the ranking of the plurality of extracted data in the first search result;
The data search system according to claim 1. The search result generating means includes
A third relevance level is obtained by weighting and adding the first relevance level and the second relevance level to the plurality of extracted data and the plurality of search target data corresponding to the plurality of extracted additional data. calculate,
The data search system according to claim 1, wherein the data search system is a data search system. 5. The apparatus according to claim 4, further comprising parameter changing means for dynamically changing a weighting value used at the time of weighting addition based on a user data selection history for the third search result output by the output means. The data retrieval system described. The parameter changing means may determine the weighting value based on a comparison between the number of times of selecting the first selection result indicated by the data selection history and the number of times of selecting the second selection result indicated by the data selection history. Change
6. A data search system according to claim 5, wherein: A search result acquisition step in which a search result acquisition means acquires a first search result indicating a plurality of extracted data extracted from a plurality of search target data by a search keyword;
An additional data storage means for storing a plurality of additional data added to the plurality of search target data;
A data search means extracts a plurality of extracted additional data related to the search keyword from the plurality of additional data, and a second search indicating the plurality of search target data to which the plurality of extracted additional data is added A data retrieval step to generate results;
The degree-of-association calculating means calculates a first degree of association with the search keyword of the plurality of extracted data based on the first search result acquired by the first search result acquisition means, and the data search A relevance calculation step of calculating a second relevance of the plurality of search target data corresponding to the plurality of extracted additional data with the search keyword based on the second search result acquired by the means;
The search result generation means adds the first relevance level and the second relevance level to the plurality of search result data and the plurality of search target data corresponding to the plurality of extracted additional data. A search result generation step for calculating the degree of relevance 3;
An output step of outputting a third search result obtained by combining the first search result and the second search result in an order based on the third degree of association;
A data search method comprising:

Images