JP5903757B2 - SEARCH DEVICE, SEARCH PROGRAM, AND SEARCH METHOD - Google Patents

Description

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

  Conventionally, as an information retrieval technique, there is a technique for retrieving information including an input keyword from an information group (for example, see Patent Document 1 below). In addition, there is a technique that generates a document vector that characterizes a document, generates a similarity between documents from among the document vectors, and searches for similar documents based on the generated similarity (see, for example, Patent Document 2 below).

Japanese Patent Laid-Open No. 9-62691 JP 2003-6214 A

  However, in the above-described conventional technology, information that does not match the search condition specified by the user is not searched from the information group to be searched. For this reason, when there is no information that matches the search condition, there is a problem that information useful for the user cannot be obtained.

  An object of the present invention is to provide a search device, a search program, and a search method capable of providing information that can be referred to by a user in information search in order to solve the above-described problems caused by the prior art.

  In order to solve the above-described problems and achieve the object, the disclosed search device, search program, and search method accept a search condition that specifies attribute information of an arbitrary attribute, and attribute information specified by the received search condition By referring to a table that defines other attribute information related to, other search conditions that specify the other attribute information are created, and based on the created other search conditions, a data group to be searched Search data from and output the search result.

  According to the search device, the search program, and the search method, it is possible to provide information that can be used as a reference for the user in information search.

1 is a block diagram illustrating a functional configuration of a search device according to a first embodiment; It is explanatory drawing which shows an example of the memory content of an attribute table. 3 is a flowchart illustrating an example of a search processing procedure of the search device according to the first embodiment; It is explanatory drawing which shows one Example of the search system concerning Embodiment 2. FIG. FIG. 3 is a block diagram illustrating a hardware configuration of a computer according to a second embodiment. It is explanatory drawing which shows an example of the memory content of an item master. It is explanatory drawing which shows an example of the memory content of a document master. It is explanatory drawing which shows an example of the memory content of a user classification table. It is explanatory drawing which shows an example of the memory content of a creation reason classification table. It is explanatory drawing which shows an example of the memory content of a creation method classification table. It is explanatory drawing which shows the specific example of a search condition input screen. It is explanatory drawing which shows the example of input of search conditions. It is a block diagram which shows the functional structure of the search device concerning Embodiment 2. It is explanatory drawing which shows an example of the memory content of a similarity definition table. It is explanatory drawing which shows an example of the memory content of a similarity table. It is explanatory drawing (the 1) which shows the example of calculation of the distance between vectors. It is explanatory drawing which shows an example of the memory content of an axis definition table. It is explanatory drawing (the 2) which shows the example of calculation of the distance between vectors. It is explanatory drawing which shows an example of the memory content of a quality table. It is explanatory drawing which shows the specific example of a search result screen. 10 is a flowchart illustrating an example of a search processing procedure of the search device according to the second exemplary embodiment; It is a flowchart which shows an example of the specific process sequence of the similarity specific process of step S2102. It is a flowchart which shows an example of the specific process sequence of the distance calculation process between vectors of step S2103. It is a flowchart which shows an example of the specific process sequence of the search process of step S2104. It is explanatory drawing which shows the specific example of a search result list.

  Exemplary embodiments of a search device, a search program, and a search method according to the present invention will be explained below in detail with reference to the accompanying drawings.

(Embodiment 1)
First, a functional configuration of the search device 100 according to the first embodiment will be described. In the first embodiment, the search condition specified by the user is converted into another search condition having a related characteristic, and the search is performed, so that the user may be helpful even if the search condition does not completely match Provide information.

  FIG. 1 is a block diagram of a functional configuration of the search device according to the first embodiment. In FIG. 1, the search device 100 is configured to include a reception unit 101, a creation unit 102, a search unit 103, and an output unit 104. Specifically, each functional unit (reception unit 101 to output unit 104) executes, for example, a program stored in a storage device such as a ROM 502, a RAM 503, a magnetic disk 505, and an optical disk 507 shown in FIG. The function is realized by making it or by the I / F 508.

  The accepting unit 101 accepts a search condition that specifies attribute information of an arbitrary attribute. Here, the search condition is a condition for searching for desired data from a data group to be searched. The search condition may be specified using, for example, characters or symbols, or may be specified from a plurality of search conditions prepared in advance.

  The search target data is data related to documents, drawings, and the like. Examples of data include specification data, design data, customer data, accounting data, and e-mail. Further, the data has attribute information of attributes that characterize the data. Data attributes include, for example, data creator, type, creation date, creation location, creation reason, creation method, and the like.

  For example, the user designates attribute information of an arbitrary attribute characterizing desired data as a search condition. Specifically, for example, the accepting unit 101 accepts attribute information of an arbitrary attribute as a search condition by a user operation input using the input device. The input device is, for example, a keyboard 510 and a mouse 511 shown in FIG.

  The creation unit 102 selects a search condition (hereinafter referred to as “second search condition”) that specifies other attribute information related to the attribute information specified by the received search condition (hereinafter referred to as “first search condition”). ). Specifically, for example, the creation unit 102 refers to the attribute table 110, so that the second search condition for designating other attribute information having the same attribute as the attribute information designated by the first search condition. Create Here, the contents stored in the attribute table 110 will be described.

  FIG. 2 is an explanatory diagram showing an example of the contents stored in the attribute table. In FIG. 2, the attribute table 110 has fields of attributes and attribute information. By setting information in each field, an attribute information group having the attributes is shown for each attribute.

  The attributes are attributes A1 to Am that characterize each data to be searched. The attribute information is specific contents of each attribute A1 to Am, for example, creator, type, creation date, creation place, creation reason, creation method, and the like. In the example of FIG. 2, attribute information T1-1 and T1-2 are shown as attribute information of the attribute A1. Further, attribute information T2-1, T2-2, T2-3, and T2-4 are shown as attribute information of the attribute A2.

  Here, it is assumed that the receiving unit 101 receives the attribute information T1-1 of the attribute A1 as the first search condition. In this case, the creation unit 102 refers to the attribute table 110 and sets a second search condition for designating another attribute information T1-2 having the same attribute A1 as the attribute information T1-1 designated by the search condition. create.

  In addition, when there are a plurality of other attribute information having the same attribute, the search unit 103 may create a second search condition for designating arbitrary attribute information from the plurality of other attribute information. . In addition, the search unit 103 may create a second search condition that specifies each of a plurality of other attribute information.

  Returning to the description of FIG. 1, the search unit 103 searches for data from the data group to be searched based on the created second search condition. Specifically, for example, the search unit 103 searches for data having other attribute information specified by the second search condition from the data group to be searched.

  Further, the search unit 103 may search for data from a data group to be searched based on the received first search condition. Specifically, for example, the search unit 103 searches for data having attribute information specified by the first search condition from the data group to be searched.

  Further, the receiving unit 101 may receive a first search condition for designating attribute information of each attribute for each attribute of each data to be searched. Specifically, for example, the receiving unit 101 receives a first search condition for designating attribute information of the attributes A1 to Am for each attribute A1 to Am included in each data.

  In this case, the creation unit 102 refers to the attribute table 110 and designates other attribute information having the same attribute as the attribute information designated in the first search condition for each of the attributes A1 to Am. Create each search condition. Then, the search unit 103 searches for data from the data group based on the second search condition created for each of the attributes A1 to Am.

  Specifically, for example, the search unit 103 searches for data having other attribute information specified by any second search condition among the second search conditions created for each of the attributes A1 to Am. May be. Moreover, the search part 103 may search the data which have all the other attribute information designated with the 2nd search condition produced for every attribute A1-Am.

  The output unit 104 outputs the searched search result. Specifically, for example, the output unit 104 may display the searched search result on the display screen. The display screen is, for example, a display 509 shown in FIG. At this time, the output unit 104 may display the search result of the first search condition and the search result of the second search condition in a state that the user can identify. Thus, by displaying the search result of the first search condition and the search result of the second search condition together, the user is beneficial to the user in addition to the information that matches the search condition specified by the user. Information that may be obtained can also be acquired.

  Further, the output unit 104 may display the search result of the first search condition with priority over the search result of the second search condition. More specifically, for example, in the search result list in which the output unit 104 lists the search results of the first and second search conditions, the search result of the first search condition is searched for the second search condition. It may be displayed at a higher rank than the result.

(Search processing procedure of search device 100)
FIG. 3 is a flowchart of an example of a search processing procedure of the search device according to the first embodiment. In the flowchart of FIG. 3, first, the receiving unit 101 determines whether or not a first search condition specifying attribute information of an arbitrary attribute has been received (step S301).

  Here, after waiting for the reception of the first search condition (step S301: No), if it is received (step S301: Yes), the creation unit 102 refers to the attribute table 110 and the first search condition is satisfied. A second search condition for designating other attribute information having the same attribute as the designated attribute information is created (step S302).

  Next, the search unit 103 searches for data from the data group to be searched based on the created second search condition (step S303). In addition, the search unit 103 searches for data from the data group to be searched based on the received first search condition (step S304). Then, the search result is output by the output unit 104 (step S305), and a series of processes according to this flowchart is terminated.

  As described above, according to the search device 100 according to the first embodiment, it is possible to perform a search by converting the search condition specified by the user into another search condition having a related feature. Thus, it is possible to provide information that may be helpful to the user even if the search condition specified by the user is not completely met.

  For example, the attribute A1 is “creator”, the attribute information T1-1 is “Taro Yamada”, and the attribute information T1-2 is “Jiro Suzuki”. “Taro Yamada” and “Jiro Suzuki” are members of the same department. Here, it is assumed that the attribute information “Taro Yamada” of the attribute “creator” is received as the first search condition.

  In this case, a second search condition for designating other attribute information “Jiro Suzuki” having the same attribute “creator” as the attribute information “Taro Yamada” is created. That is, a document created by “Jiro Suzuki” has a higher possibility of information that is helpful to the user than a document created by a person belonging to a different department from “Taro Yamada”.

  Therefore, in the first embodiment, a second search condition for designating “Jiro Suzuki” belonging to the same department as “Taro Yamada” is created. As a result, for example, documents created by “Jiro Suzuki” belonging to the same department other than documents created by “Taro Yamada” can be searched, and the search conditions specified by the user are not completely matched. Can also provide information that may be helpful to the user.

(Embodiment 2)
Next, a search system 400 according to the second embodiment will be described. In the second embodiment, a PDM (Product Data Management) system will be described as an example of the search system 400. In addition, about the same location as the location demonstrated in Embodiment 1, the same code | symbol is attached | subjected and description is abbreviate | omitted.

  FIG. 4 is an explanatory diagram of an example of the search system according to the second embodiment. In FIG. 4, the search system 400 includes a search device 401 and a plurality of client terminals 402. In the search system 400, a search device 401 and a plurality of client terminals 402 are connected to each other via a network 410 such as the Internet, a LAN (Local Area Network), or a WAN (Wide Area Network).

  Here, the search system 400 is, for example, a PDM system that centralizes and manages information related to design / development in an industrial product development process, thereby improving the efficiency of the development process and shortening the period. Specifically, for example, the search system 400 manages drawing data such as CAD (Computer Aided Design) data, document data such as specification data, and component data included in a product.

  The search device 401 includes a database that stores a data group to be searched, and is a computer that searches for data from the database. The database is, for example, the item master 403 and the document master 404. Specifically, for example, the search device 401 receives a search condition from the client terminal 402 via the network 410 and searches for data that matches the search condition from a data group to be searched. Detailed descriptions of the item master 403 and the document master 404 will be described later with reference to FIGS.

  The client terminal 402 is a computer used by a user who uses the search system 400. The client terminal 402 receives input of search conditions from the user and transmits the search conditions to the search device 401. As a result, the client terminal 402 receives the search result from the search device 401 and displays the search result on the display screen.

  In the following description, a data group to be searched is expressed as “data group D1 to Dn”, and arbitrary data among the data groups D1 to Dn is expressed as “data Di” (i = 1, 2,..., n).

  Next, the hardware configuration of the search device 401 and the client terminal 402 shown in FIG. 4 will be described with reference to FIG. In the description of FIG. 5, the search device 401 and the client terminal 402 illustrated in FIG. 4 are simply referred to as “computers”.

(Computer hardware configuration)
FIG. 5 is a block diagram of a hardware configuration of the computer according to the second embodiment. In FIG. 5, the computer includes a CPU (Central Processing Unit) 501, a ROM (Read-Only Memory) 502, a RAM (Random Access Memory) 503, a magnetic disk drive 504, a magnetic disk 505, and an optical disk drive 506. , An optical disk 507, an I / F (Interface) 508, a display 509, a keyboard 510, and a mouse 511. Each component is connected by a bus 500.

  Here, the CPU 501 governs overall control of the computer. The ROM 502 stores a program such as a boot program. The RAM 503 is used as a work area for the CPU 501. The magnetic disk drive 504 controls reading / writing of data with respect to the magnetic disk 505 according to the control of the CPU 501. The magnetic disk 505 stores data written under the control of the magnetic disk drive 504.

  The optical disk drive 506 controls the reading / writing of the data with respect to the optical disk 507 according to control of CPU501. The optical disk 507 stores data written under the control of the optical disk drive 506, or causes the computer to read data stored on the optical disk 507.

  The I / F 508 is connected to a network 410 such as a LAN, a WAN, or the Internet through a communication line, and is connected to another device via the network 410. The I / F 508 controls an internal interface with the network 410 and controls input / output of data from an external device. For example, a modem or a LAN adapter may be employed as the I / F 508.

  A display 509 displays data such as a document, an image, and function information as well as a cursor, an icon, or a tool box. As the display 509, for example, a CRT, a TFT liquid crystal display, a plasma display, or the like can be adopted.

  The keyboard 510 includes keys for inputting characters, numbers, various instructions, and the like, and inputs data. Moreover, a touch panel type input pad or a numeric keypad may be used. The mouse 511 moves the cursor, selects a range, moves the window, changes the size, and the like. A trackball or a joystick may be used as long as they have the same function as a pointing device.

  Note that the search device 401 may not include the display 509, the keyboard 510, and the mouse 511 described above. The client terminal 402 may include a scanner that reads an image and captures image data, and a printer that prints image data and document data.

(Memory contents such as various tables)
Hereinafter, storage contents of various tables and the like included in the search device 401 will be described with reference to FIGS. The various tables are realized by a storage device such as the ROM 502, the RAM 503, the magnetic disk 505, and the optical disk 507 of the search device 401 shown in FIG.

<Storage contents of item master 403>
FIG. 6 is an explanatory diagram of an example of the contents stored in the item master. In FIG. 6, the item master 403 has fields of data ID, number, type, creation date, creator, creation reason, creation method, creation location, and quality. By setting information in each field, item information 600-1 to 600-n related to product items is stored as a record.

  Here, the item information 600-1 to 600-n is a data group D1 to Dn to be searched. The content of each field is attribute information of each attribute characterizing the item information 600-1 to 600-n. Specifically, the data ID is an identifier of each item information 600-1 to 600-n. The number is an identifier of a product to be procured, produced, stored and sold.

  The type is the type of data Di. As the type, for example, a model, a unit, a part, and the like that specify the type of product are set. The creation date is the creation date of the data Di. The creator is the creator of the data Di. The reason for creation is the reason that created the data Di. The reasons for creation include, for example, new product creation, product correction, product failure handling, product performance improvement, and product cost reduction.

  The creation method is a method of creating data Di. Examples of the creation method include new creation, diversion of other data, import from another system, and automatic creation by the system. The creation location is a location where the data Di is created. As the creation location, for example, the IP (Internet Protocol) address of the client terminal 402 used by the creator of the data Di is set. The quality is an index value that represents the quality of the product. Here, the quality is expressed by the number of complaints from customers, and the lower the value, the higher the quality.

  Taking the item information 600-1 as an example, the data ID is “D1”, the number is “Part1”, the type is “model”, the creation date is “2010/5/1”, the creator is “U1”, and the reason for creation Is “R1”, the creation method is “M1”, and the creation location is “10.67. xxx. xx ”and the quality is“ 0 ”.

<Storage contents of document master 404>
FIG. 7 is an explanatory diagram showing an example of the contents stored in the document master. In FIG. 7, the document master 404 has fields of data ID, number, type, creation date, creator, creation reason, creation method, creation location, and quality. By setting information in each field, document information 700-1 to 700-n related to products is stored as records.

  Here, the document information 700-1 to 700-n is a data group D1 to Dn to be searched. The content of each field is attribute information of each attribute characterizing the document information 700-1 to 700-n. Specifically, the data ID is a record identifier of each document information 700-1 to 700-n. The number is an identifier of an electronic document relating to a product to be procured, produced, stored or sold. Examples of electronic documents related to products include CAD data, specification data, and contract data.

  The type is the type of data Di. As the type, for example, a parts table, a 2D drawing, a 3D drawing, and a design change notification table indicating the type of electronic document relating to the product are set. The creation date is the creation date of the data Di. The creator is the creator of the data Di. The reason for creation is the reason that created the data Di. The reasons for creation include, for example, new product creation, product correction, product failure handling, product performance improvement, and product cost reduction.

  The creation method is a method of creating data Di. Examples of the creation method include new creation, diversion of other data, import from another system, and automatic creation by the system. The creation location is a location where the data Di is created. As the creation location, for example, the IP address of the client terminal 402 used by the creator of the data Di is set. The quality is an index value that represents the quality of the product. Here, the quality is expressed by the number of complaints from customers, and the lower the value, the higher the quality.

  Taking document information 700-1 as an example, the data ID is “D1”, the product number is “Doc1”, the type is “BOM”, the creation date is “2010/2/1”, the creator is “U1”, creation The reason is “R10”, the creation method is “M4”, and the creation location is “10.67. xxx. xx ”and the quality is“ 0 ”.

<Storage contents of user classification table 800>
FIG. 8 is an explanatory diagram of an example of the contents stored in the user classification table. In FIG. 8, a user classification table 800 has fields of group ID and user ID, and group information (for example, group information 800-1 and 800-2) is stored as a record by setting information in each field. doing.

  Here, the group ID is an identifier of a group that classifies a user group that uses the search system 400. The user ID is an identifier of a user who uses the search system 400. As an example of classification, a user group using the search system 400 is classified by, for example, organizational classification such as a department in a company, or geographical classification such as a head office or a branch.

  That is, the group information in the user classification table 800 represents, for example, a set of users who are likely to have similar problems because the business contents are similar. Taking the group information 800-1 as an example, the users belonging to the group “UG1” are “U1, U2, U3”. This indicates that the users U1, U2, and U3 have similar business contents and are likely to have similar tasks they face.

<Storage contents of creation reason classification table 900>
FIG. 9 is an explanatory diagram of an example of the contents stored in the creation reason classification table. In FIG. 9, a creation reason classification table 900 has fields for group ID and creation reason. By setting information in each field, group information (for example, group information 900-1 to 900-3) is used as a record. I remember it.

  Here, the group ID is an identifier of a group that classifies the reason for creating each data Di in the item master 403 and the document master 404 shown in FIGS. The reason for creation is the reason that created each data Di. As an example of classification, the reasons for creating "product correction" and "product failure handling" are classified into the same group, and the reasons for creating "product performance improvement" and "product cost reduction" are classified into the same group. Is done.

  That is, the group information in the creation reason classification table 900 represents a set of creation reasons that are similar in reason for creating each data Di. Taking the group information 900-2 as an example, the reason for creation belonging to the group “RG2” is “R2, R3”. This indicates that the creation reasons R2 and R3 are similar in the reason for creating the data Di.

<Storage contents of creation method classification table 1000>
FIG. 10 is an explanatory diagram showing an example of the contents stored in the creation method classification table. In FIG. 10, a creation method classification table 1000 has fields for group ID and creation method. By setting information in each field, group information (for example, group information 1000-1 to 1000-3) is used as a record. I remember it.

  Here, the group ID is an identifier of a group for classifying the creation method of each data Di in the item master 403 and the document master 404 shown in FIGS. The creation method is a creation method of each data Di. As an example of classification, “new creation” and “diversion of other products” creation methods are classified into the same group.

  That is, the group information in the creation method classification table 1000 represents a set of creation methods in which the methods for creating each data Di are similar. Taking the group information 1000-1 as an example, the creation method belonging to the group “MG1” is “M1, M2”. This indicates that the creation methods M1 and M2 are similar in the method of creating the data Di.

  In the above description, the search device 401 is provided with various tables and the like, but is not limited thereto. For example, the search system 400 may include a database server including various tables. In this case, the search device 401 refers to stored contents such as various tables by accessing the database server.

(Specific example of search condition input screen)
Next, a search condition input screen 1100 displayed on the display 509 of the client terminal 402 will be described with reference to FIG. The search condition input screen 1100 is displayed on the display 509 of the client terminal 402 when a user using the search system 400 inputs search conditions for searching for desired data.

  FIG. 11 is an explanatory diagram showing a specific example of the search condition input screen. In FIG. 11, the search condition input screen 1100 includes boxes 1101 to 1107 for specifying a search condition, a search button B1, and a clear button B2.

  On the search condition input screen 1100, the search condition for each attribute is input by moving the cursor C and clicking each box 1101 to 1107 by the user's operation input using the keyboard 510 or the mouse 511 shown in FIG. can do.

  Specifically, when the cursor C is moved and the box 1101 is clicked, a product number or an electronic document number related to the product can be input as a search condition. When the cursor C is moved and the box 1102 is clicked, the type of data Di to be searched can be input as a search condition.

  When the cursor C is moved and the box 1103 is clicked, the creator of the data Di to be searched can be input as a search condition. When the cursor C is moved and the box 1104 is clicked, the creation date of the data Di to be searched can be input as a search condition. When the cursor C is moved and the box 1105 is clicked, the reason for creating the data Di to be searched can be input as a search condition.

  When the cursor C is moved and the box 1106 is clicked, a method for creating data Di to be searched can be input as a search condition. When the cursor C is moved and the box 1107 is clicked, the creation location of the data Di to be searched can be input as a search condition.

  Further, when the input of the search condition is completed, when the cursor C is moved and the search button B1 is clicked, a search based on the input search condition is performed, and the search result is displayed on the display 509. If the cursor C is moved and the clear button B2 is clicked, the already entered search condition is reset.

  FIG. 12 is an explanatory diagram showing an example of input of search conditions. In FIG. 12, “parts” is input in the box 1102 of the search condition input screen 1100, “U1” is input in the box 1103, and “2010/6/1” is input in the box 1104.

  Here, when the search button B1 is clicked, the type “part”, the creator “U1”, and the creation date “2010/6/1” of the data Di to be searched are specified as the search condition 1200, and the client terminal 402 The search condition 1200 is transmitted from the search device 401 to the search device 401.

  In the following description, when the product number (Part #) and the product type (model, unit, etc.) are designated as the search conditions, the search device 401 specifies the item information 600-1 to 600-1 in the item master 403. A search is performed with 600-n as a search target (# = 1, 2,...). In addition, when the electronic document number (Doc #) or electronic document type (parts table, 2D drawing, etc.) is designated as a search condition, the search device 401 specifies document information 700-1 to 700-700 in the document master 404. -N is searched for.

(Functional configuration of search device 401)
FIG. 13 is a block diagram of a functional configuration of the search device according to the second embodiment. In FIG. 13, the search device 401 includes a receiving unit 101, a search unit 103, an output unit 104, a specifying unit 1301, a first calculation unit 1302, and a second calculation unit 1303. . Specifically, for example, each function unit causes the CPU 501 to execute a program stored in a storage device such as the ROM 502, the RAM 503, the magnetic disk 505, and the optical disk 507 of the search device 401, or by the I / F 508. Realize its function. Further, the processing results of the respective functional units are stored in a storage device such as the RAM 503, the magnetic disk 505, and the optical disk 507, unless otherwise specified.

  First, the reception unit 101 receives a search condition. Here, the search condition is, for example, the search condition 1200 shown in FIG. 12, and designates attribute information of attributes that characterize the data Di to be searched. Specifically, for example, the receiving unit 101 receives the search condition 1200 from the client terminal 402 via the network 410. Further, the receiving unit 101 may receive a search condition directly input to the search device 401 by a user operation input using the keyboard 510 or the mouse 511 of the search device 401.

  For each attribute that characterizes each data Di to be searched, the specifying unit 1301 specifies the similarity between the attribute information of the attribute specified by the received search condition and the attribute information of the attribute of each data Di. Specifically, for example, the specifying unit 1301 specifies the similarity for each attribute of each data Di by referring to a table that defines the similarity between attribute information having the same attribute. The specific processing content of the specifying unit 1301 will be described later with reference to FIG.

  The first calculation unit 1302 multi-dimensionalizes the search condition and each data Di using the attribute specified by the search condition as a vector element based on the similarity for each attribute of each specified data Di. A vector distance Li between the search condition and each data Di is calculated. The inter-vector distance Li represents the similarity between the search condition and the data Di. The data Di having a smaller inter-vector distance Li is more similar (applicable) to the search condition. Note that the specific processing content of the first calculation unit 1302 will be described later with reference to FIG.

  The search unit 103 searches for data from the data groups D1 to Dn to be searched based on the calculated intervector distance Li of each data Di. Specifically, for example, the search unit 103 may search the data Di having the smallest inter-vector distance Li from the data groups D1 to Dn. The search process by the search unit 103 is repeatedly performed until, for example, a preset number (for example, m) of search results are obtained.

  The output unit 104 outputs the searched search result. Specifically, for example, the output unit 104 transmits the searched search result to the client terminal 402 via the network 410. As a result, the search result is displayed on the display 509 of the client terminal 402. Note that an example of a screen displayed on the display 509 of the client terminal 402 will be described later with reference to FIG.

  Other output formats of the output unit 104 include, for example, display on the display 509 of the search device 401 and print output to a printer (not shown). The output unit 104 may store the search results in a storage area such as the RAM 503, the magnetic disk 505, and the optical disk 507.

<Specific Processing Contents of the Identification Unit 1301>
Here, an example of specific processing contents of the specifying unit 1301 will be described. First, the similarity definition table 1400 that defines the similarity between attribute information having the same attribute will be described. The similarity definition table 1400 is stored in a storage device such as the ROM 502, the RAM 503, the magnetic disk 505, and the optical disk 507, for example.

  FIG. 14 is an explanatory diagram of an example of the contents stored in the similarity definition table. In FIG. 14, the similarity definition table 1400 has attributes, relationships, and similarity fields. By setting information in each field, the similarity definition information 1400-1 to 1400-6 is stored as a record. Yes.

  Here, the attribute is an attribute that characterizes the data Di to be searched. The relationship is a relationship between attribute information having the same attribute. The similarity is the similarity between attribute information having the same attribute. The degree of similarity indicates that the greater the value, the greater the degree of similarity. The minimum value of the similarity is “0”, and the maximum value is “1”.

  Taking similarity definition information 1400-1 as an example, the similarity between attribute information having the attribute of “creator” is shown. Specifically, if the relationship between attribute information is “same user”, similarity is “1”, if “same group”, similarity is “0.6”, and similarity is “other”. Is “0”. The specifying unit 1301 specifies the similarity for each attribute of each data Di to be searched by referring to the similarity definition table 1400, for example.

  Here, as an example of the specific processing contents of the specifying unit 1301, the item information 600-1 in the item master 403 is taken as an example of the data Di to be searched, and the search condition 1200 is received as an example. explain.

In the following description, a plurality of attributes specified by the search condition are expressed as “attributes A1 to AK”, and an arbitrary attribute among the attributes A1 to AK is expressed as “attribute Ak” (k = 1, 2,... , K). Further, the similarity between the attribute information of the attribute Ak between the search condition and the data Di is expressed as “similarity R _i (k)”.

  First, the specifying unit 1301 specifies the attributes A1 to AK specified by the search condition 1200. In the search condition 1200, for example, the attribute A1 is “type”, the attribute A2 is “creator”, and the attribute A3 is “creation date”. Next, the specifying unit 1301 pays attention to the attribute A1 and specifies the relationship between the attribute information of the attribute A1 specified by the search condition 1200 and the attribute information of the attribute A1 of the item information 600-1.

  Here, the attribute information of the attribute A1 specified by the search condition 1200 is “part”, and the attribute information of the attribute A1 of the item information 600-1 is “model”. Therefore, the relationship between the attribute information of the attribute A1 specified by the search condition 1200 and the attribute information of the attribute A1 of the item information 600-1 is “others”.

Thereafter, the identifying unit 1301 refers to the similarity definition table 1400 to identify the similarity having the attribute “kind” and the relationship “other”. Here, the degree of similarity is “0”. As a result, the specifying unit 1301 sets the similarity R ₁ (1) between the attribute information of the attribute A1 specified by the search condition 1200 and the attribute information of the attribute A1 of the item information 600-1 to “R ₁ (1) = 0. ] Specified.

  Next, the specifying unit 1301 pays attention to the attribute A2, and specifies the relationship between the attribute information of the attribute A2 specified by the search condition 1200 and the attribute information of the attribute A2 of the item information 600-1. Here, the attribute information of the attribute A2 designated by the search condition 1200 is “U1”, and the attribute information of the attribute A2 of the item information 600-1 is “U1”. Therefore, the relationship between the attribute information of the attribute A2 specified by the search condition 1200 and the attribute information of the attribute A2 of the item information 600-1 is “same user”.

Thereafter, the identifying unit 1301 refers to the similarity definition table 1400 and identifies the similarity with the attribute “creator” and the relationship “same user”. Here, the similarity is “1”. As a result, the specifying unit 1301 sets the similarity R ₁ (2) between the attribute information of the attribute A2 specified by the search condition 1200 and the attribute information of the attribute A2 of the item information 600-1 to “R ₁ (2) = 1. ] Specified.

  Then, the specifying unit 1301 pays attention to the attribute A3 and specifies the relationship between the attribute information of the attribute A3 specified by the search condition 1200 and the attribute information of the attribute A3 of the item information 600-1. Here, the attribute information of the attribute A3 designated by the search condition 1200 is “2010/6/1”, and the attribute information of the attribute A3 of the item information 600-1 is “2010/5/1”. Therefore, the relationship between the attribute information of the attribute A3 specified by the search condition 1200 and the attribute information of the attribute A3 of the item information 600-1 is “within 2 months”.

Thereafter, the identifying unit 1301 refers to the similarity definition table 1400 and identifies the similarity having the attribute “creation date” and the relationship “within two months”. Here, the similarity is “0.7”. As a result, the similarity R ₁ (3) between the attribute information of the attribute A3 specified by the search condition 1200 and the attribute information of the attribute A3 of the item information 600-1 is specified as “R ₁ (3) = 0.7”. To do.

Regarding the attribute “creator” In addition, for the attribute “creator”, the specifying unit 1301 has a relationship between the attribute information specified in the search condition and the attribute information of the data Di to be retrieved is not “same user”. The relationship between attribute information is specified with reference to the user classification table 800 shown in FIG.

  Specifically, for example, first, the specifying unit 1301 refers to the user classification table 800 and specifies the group UG to which the user indicated by the attribute information specified by the search condition belongs. Next, the specifying unit 1301 determines whether or not the user indicated by the attribute information of the data Di to be searched belongs to the specified group UG. Here, when belonging to the group UG, the specifying unit 1301 specifies the relationship “same group” between the attribute information. On the other hand, when it does not belong to the group UG, the identifying unit 1301 identifies the relationship “others” between the attribute information.

Regarding the attribute “reason for creation” In addition, for the attribute “reason for creation”, the specifying unit 1301 has a relationship between the attribute information specified in the search condition and the attribute information of the data Di to be searched is not “same reason”. Referring to the creation reason classification table 900 shown in FIG. 9, the relationship between the attribute information is specified.

  Specifically, for example, first, the identifying unit 1301 refers to the creation reason classification table 900 and identifies the group RG to which the creation reason indicated by the attribute information specified by the search condition belongs. Next, the specifying unit 1301 determines whether or not the creation reason indicated by the attribute information of the data Di to be searched belongs to the specified group RG. Here, when belonging to the group RG, the identifying unit 1301 identifies the relationship “same group” between the attribute information. On the other hand, when it does not belong to the group RG, the identifying unit 1301 identifies the relationship “others” between the attribute information.

Attribute “Creation Method” In addition, for the attribute “creation method”, the specifying unit 1301 determines that the relationship between the attribute information specified by the search condition and the attribute information of the data Di to be retrieved is not “same method”. The relationship between attribute information is specified with reference to the creation method classification table 1000 shown in FIG.

  Specifically, for example, the identifying unit 1301 first refers to the creation method classification table 1000 to identify the group MG to which the creation method indicated by the attribute information specified by the search condition belongs. Next, the specifying unit 1301 determines whether the creation method indicated by the attribute information of the data Di to be searched belongs to the specified group MG. Here, when belonging to the group MG, the specifying unit 1301 specifies the relationship “same group” between the attribute information. On the other hand, when it does not belong to the group MG, the identifying unit 1301 identifies the relationship “others” between the attribute information.

The similarity R _i (k) for each attribute Ak that characterizes each identified data Di is stored, for example, in the similarity table 1500 shown in FIG. The similarity table 1500 is realized by a storage device such as the ROM 502, the RAM 503, the magnetic disk 505, and the optical disk 507, for example.

FIG. 15 is an explanatory diagram of an example of the contents stored in the similarity table. In FIG. 15, the similarity table 1500 has a similarity R _i (k) for each attribute Ak of each data Di to be searched.

In the example of the search condition 1200 and the item information 600-1, the similarity R ₁ (1) of the attribute A1 of the data D1 in the similarity table 1500 is “R ₁ (1) = 0”. The similarity R ₁ (2) of the attribute A2 of the data D1 is “R ₁ (2) = 1”. The similarity R ₁ (3) of the attribute A3 of the data D1 is “R ₁ (3) = 0.7”.

<Specific Processing Contents of First Calculation Unit 1302>
Next, an example of specific processing contents of the first calculation unit 1302 that calculates the inter-vector distance Li between the search condition and the data Di will be described. Here, the search condition 1200 shown in FIG. 12 will be described as the search condition, and the item information 600-1 in the item master 403 will be described as an example of the data Di.

  FIG. 16 is an explanatory diagram (part 1) of an example of calculating the distance between vectors. (I) The first calculation unit 1302 refers to the axis definition table 1700 illustrated in FIG. 17 to identify the axis corresponding to each attribute Ak specified by the search condition. Here, the axis definition table 1700 will be described. The axis definition table 1700 is stored in a storage device such as the ROM 502, the RAM 503, the magnetic disk 505, and the optical disk 507, for example.

  FIG. 17 is an explanatory diagram of an example of the contents stored in the axis definition table. In FIG. 17, the axis definition table 1700 has fields of axis, attribute, and importance. By setting information in each field, axis definition information 1700-1 to 1700-7 is stored as a record.

  Here, the axis is a straight line indicating a vector component of each dimension in the orthogonal coordinate system. The attribute is an attribute that characterizes the data Di to be searched. The importance is an index value representing the importance of each attribute. Here, the smaller the value, the higher the importance. The importance of each attribute can be set arbitrarily. Specifically, for example, as the importance of each attribute, a high importance is set for the attribute that should be prioritized in the similarity determination with the search condition. For example, when it is desired to search for data Di whose “type” is similar to the search condition from “creator”, the importance of the attribute “type” is set higher than that of the attribute “creator”.

  Each axis represents each factor related to decision making called 5W1H (Who, When, What, Where, Why, How) and product quality. Specifically, the Who axis corresponds to the attribute “creator” indicating who created the data Di to be searched.

  The When axis corresponds to an attribute “creation date” indicating when the data Di to be searched is created. The What axis corresponds to the attributes “type” and “number” that indicate what the data Di to be searched relates to. The Where axis corresponds to an attribute “creation location” that indicates where the data Di to be searched is created.

  The Why axis corresponds to the attribute “reason for creation” indicating why the data Di to be searched is created. The How axis corresponds to an attribute “creation method” indicating how data Di to be searched is created. The Quality axis corresponds to the attribute “quality” representing the quality of the product.

  In the example of the search condition 1200, the first calculation unit 1302 refers to the axis definition table 1700, and specifies the axis “What” corresponding to the attribute A1 “type” specified in the search condition 1200. Also, the first calculation unit 1302 refers to the axis definition table 1700 and identifies the axis “Who” corresponding to the attribute A2 “creator” specified by the search condition 1200.

  Also, the first calculation unit 1302 refers to the axis definition table 1700 and identifies the axis “When” corresponding to the attribute A3 “creation date” specified in the search condition 1200. Accordingly, a three-dimensional orthogonal coordinate system 1600 is defined by the What axis, the Who axis, and the When axis corresponding to the attributes A1 to A3.

  (Ii) The first calculation unit 1302 obtains a vector (hereinafter referred to as “reference vector SV”) representing a search condition in the orthogonal coordinate system 1600 in which the axis corresponding to each specified attribute Ak is radially extended from the origin. create. Specifically, for example, the first calculation unit 1302 creates a reference vector SV in which the component of the axis corresponding to each attribute Ak is “1”.

  That is, when the reference vector SV is displayed as a component, it can be expressed as the following equation (1). However, SV represents the component display of the reference vector SV. A1 to AK represent axis components corresponding to the attributes A1 to AK.

    SV = (A1, A2,..., AK) = (1,1,..., 1) (1)

  In the example of the search condition 1200, the first calculation unit 1302 creates a reference vector SV in which the components of the What axis, the Who axis, and the When axis corresponding to each attribute A1 to A3 are “1” in the orthogonal coordinate system 1600. To do.

(Iii) The first calculation unit 1302 creates a vector (hereinafter referred to as “comparison vector Vi”) representing the data Di to be searched. Specifically, for example, the first calculation unit 1302 creates a comparison vector Vi in which the axis component corresponding to each attribute Ak is the similarity R _i (k) of the attribute Ak.

That is, when the comparison vector Vi is displayed as a component, it can be expressed as the following equation (2). However, Vi represents the component display of the comparison vector Vi. A1 to AK represent axis components corresponding to the attributes A1 to AK. R _i (1) to R _i (K) are similarity degrees between attribute information for each attribute A1 to AK of the search condition and the data Di.

Vi = (A1, A2,..., AK) = (R _i (1), R _i (2),..., R _i (K))
... (2)

  In the example of the item information 600-1, the first calculation unit 1302 sets “What”, “Who” axis, and “When” axis components corresponding to the respective attributes A1 to A3 to “0” and “1” in the orthogonal coordinate system 1600, respectively. And a comparison vector V1 of “0.7” is created.

  (Iv) The first calculation unit 1302 calculates an inter-vector distance Li between the created reference vector SV and the comparison vector Vi. Specifically, for example, the first calculation unit 1302 can calculate the inter-vector distance Li between the reference vector SV and the comparison vector Vi using the following equation (3). Here, Li is the inter-vector distance between the reference vector SV and the comparison vector Vi.

  In the example of the search condition 1200 and the item information 600-1, the inter-vector distance L1 between the reference vector SV and the comparison vector V1 is “L1≈1.044”. The processes (iii) and (iv) are performed in the same manner for the item information 600-2 to 600-n in the item master 403, for example.

  As a result, vector distances L1 to Ln between the search condition 1200 and the item information 600-1 to 600-n in the item master 403 are calculated. Then, the search unit 103 searches for item information that meets the search condition 1200 from the item master 403 based on the calculated inter-vector distances L1 to Ln.

Thereby, even if the search condition specified by the user does not completely match, it is possible to search for data Di that may be helpful to the user. At this time, by obtaining a similarity R _i (k) by making each factor related to human decision making into a vector element and obtaining a similarity R _i (k), a document created by an author who is in the same situation as the searcher Drawings and the like can be provided to searchers.

Priorities among data having the same inter-vector distance When the search unit 103 searches for data Di having the smallest inter-vector distance Li, there may be a plurality of data having the smallest inter-vector distance. In this case, the search unit 103 determines which data is most suitable for the search condition from among a plurality of data having the smallest inter-vector distance (hereinafter referred to as “data D [1] to D [T]”). Can not do it.

  Therefore, among the data D [1] to D [T], data that is more similar to the search condition for the specific attribute (hereinafter referred to as “attribute AP”) of the attributes A1 to AK specified by the search condition. Treat as data that best meets the search criteria. Hereinafter, an example of specific processing contents for determining priority between data having the same vector distance will be described.

  FIG. 18 is an explanatory diagram (part 2) of an example of calculating the distance between vectors. In FIG. 18, (i) a comparison vector Vx and a comparison vector Vy having the same inter-vector distance from the reference vector SV are shown. The comparison vector Vx is a vector representing the data Dx. The comparison vector Vy is a vector representing the data Dy.

  (Ii) The second calculation unit 1303 vectorizes the axis representing the specific attribute AP. Specifically, for example, the second calculation unit 1303 sets the axis component representing the specific attribute AP among the attributes A1 to AK to “1”, and sets the axis component representing the other attributes to “0”. To create a vector (hereinafter referred to as “axis vector PV”).

  Here, the specific attribute AP is, for example, the attribute having the highest importance in the data search among the attributes A1 to AK. In the example of the search condition 1200, among the attributes A1 to A3, the attribute A1 “kind” has the highest importance, so the specific attribute AP is the attribute A1. Therefore, the second calculation unit 1303 creates an axis vector PV in which the axis component representing the attribute A1 is “1” and the axis component representing the attributes A2 and A3 is “0”.

  (Iii) The second calculation unit 1303 calculates the inter-vector distance L between the comparison vector V [t] representing each data D [t] of the data D [1] to D [T] and the created axis vector PV. [T] is calculated. Specifically, for example, the second calculation unit 1303 can calculate the inter-vector distance L [t] between the comparison vector V [t] and the axis vector PV using the following equation (4).

Here, L [t] is the inter-vector distance between the comparison vector V [t] and the axis vector PV. P (k) is an axis component (1 or 0) corresponding to the attribute Ak of the axis vector PV. R _[t] (k) is an axis component (similarity) corresponding to the attribute Ak of the comparison vector V [t].

  In the example of the search condition 1200 and the data Dx, Dy, the second calculation unit 1303 uses the above equation (4) to calculate the comparison vectors Vx, Vy representing the data Dx, Dy and the created axis vector PV. Inter-vector distances L [x] and L [y] are calculated. As a result, a vector distance L [t] between the search condition and each data D [t] is calculated.

  The search unit 103 searches the data D [1] to D [T] for data that meets the search condition based on the calculated inter-vector distance L [t]. For example, when the magnitude relationship between the vector distances L [x] and L [y] is “L [x]> L [y]”, the search unit 103 sets the data having the smallest vector distance among the data Dx and Dy. Search for Dy.

  As a result, even when there are a plurality of pieces of data having the minimum distance between vectors, the data Di to be searched can be uniquely specified. Specifically, data D [t] that is more similar to the search condition can be searched for a specific attribute AP that should be prioritized in the similarity determination with the search condition.

Addition of Axis Representing Product Quality Next, a case where an axis corresponding to the product quality is added in addition to the axis corresponding to the attributes A1 to AK specified by the search condition will be described.

  The first calculation unit 1302 creates a reference vector SV representing a search condition in an orthogonal coordinate system in which an axis corresponding to each attribute Ak specified by the search condition and an axis corresponding to quality are radially extended from the origin. Specifically, for example, the first calculation unit 1302 creates a reference vector SV in which the component of each axis is “1”. That is, when the reference vector SV is displayed as a component, it can be expressed as the following equation (5). However, Q represents the component of the axis | shaft corresponding to quality.

    SV = (A1, A2,..., AK, Q) = (1,1,..., 1,1) (5)

Next, the first calculation unit 1302 creates a comparison vector Vi representing the data Di to be searched. Here, the component of the axis corresponding to each attribute Ak of the comparison vector Vi is the similarity R _i (k) of each attribute Ak. Further, the axis component corresponding to the quality of the comparison vector Vi is specified by referring to, for example, the quality table 1900 shown in FIG. The quality table 1900 is stored in a storage device such as the ROM 502, the RAM 503, the magnetic disk 505, and the optical disk 507, for example.

  FIG. 19 is an explanatory diagram of an example of the contents stored in the quality table. In FIG. 19, the quality table 1900 has coordinate values for each number of complaints. The number of complaints is an index value representing the quality of the product. The coordinate value is an axis component corresponding to the quality of the comparison vector Vi. Here, the smaller the number of claims, that is, the higher the product quality, the larger the axis component corresponding to the quality of the comparison vector Vi.

Taking the item information 600-1 as an example, since the quality is “0”, the component of the axis corresponding to the quality of the comparison vector V1 is “1”. That is, when the comparison vector Vi is displayed as a component, it can be expressed as the following equation (6). Here, q _i is an axis component corresponding to the quality of the comparison vector Vi.

Vi = (A1, A2, ..., AK, Q)
= (R _i (1), R _i (2),..., R _i (K), q _i ) (6)

  Then, the first calculation unit 1302 calculates an inter-vector distance Li between the created reference vector SV and the comparison vector Vi. Specifically, for example, the first calculation unit 1302 can calculate the inter-vector distance Li between the reference vector SV and the comparison vector Vi using the following equation (7).

  Based on the calculated inter-vector distance Li, the search unit 103 searches for data that meets the search condition from the data groups D1 to Dn. Specifically, for example, the search unit 103 searches the data Di having the smallest inter-vector distance Li from the data groups D1 to Dn.

  Thus, by setting the axis component corresponding to the quality of the comparison vector Vi representing the data Di relating to the high-quality product to a large value, the inter-vector distance Li of the data Di relating to the high-quality product is reduced. For this reason, it becomes easy to search the data Di regarding the high quality product, and as a result, it is possible to provide the user with drawings and documents regarding the high quality product.

Cumulative similarity In the above description, the case where data matching the search condition is searched based on the inter-vector distance Li between the search condition and the data Di, but the similarity R _i for each attribute Ak of the data Di is described. Data may be searched based on a value obtained by simply accumulating (k).

Specifically, for example, the first calculation unit 1302 accumulates the similarity R _i (k) for each attribute Ak of the data Di (hereinafter, “accumulated similarity R” for each data Di to be searched). _i ”). The cumulative similarity R _i represents the similarity between the search condition and the data Di. The data Di having a larger cumulative similarity R _i is more similar (adapted) to the search condition.

Then, the search unit 103 may search for data from the data groups D1 to Dn to be searched based on the calculated cumulative similarity R _i of each data Di. Specifically, for example, the search unit 103 may search for data having the maximum cumulative similarity R _i from the data groups D1 to Dn.

Thereby, even if the search condition specified by the user does not completely match, it is possible to search for data Di that may be helpful to the user. In the search process based on the cumulative similarity R _i , the concept of direction is lost as compared with the search process based on the inter-vector distance Li described above, and therefore there is a possibility that the accuracy of similarity determination with the search condition may be reduced. Thus, the load on the calculation process of the first calculation unit 1302 can be reduced.

(Specific example of search result screen)
FIG. 20 is an explanatory diagram illustrating a specific example of a search result screen. In FIG. 20, item information 600-1, 600-3, and 600-4 are displayed on the search result screen 2000 as search results that match the search conditions specified by the user. This search result screen 2000 is displayed on the display 509 of the client terminal 402, for example.

  On the search result screen 2000, when a user inputs an operation using the keyboard 510 or the mouse 511 of the client terminal 402 to move the cursor C and click on an arbitrary search result, document information related to the search result is displayed. You may decide to do it. For example, when the cursor C is moved and the item information 600-1 is clicked, the document information 700-1 in the document master 404 is displayed.

  Information about the correspondence between the item information and the document information is stored in a storage device such as the ROM 502, the RAM 503, the magnetic disk 505, and the optical disk 507, for example. Specifically, for example, the correspondence between item information and document information is set in advance when each item information and document information is registered in the item master 403 and document master 404.

(Search processing procedure of search device 401)
FIG. 21 is a flowchart of an example of a search processing procedure of the search device according to the second embodiment. Here, a case where the data group to be searched is the item information 600-1 to 600-n in the item master 403 will be described as an example.

  In the flowchart of FIG. 21, first, the receiving unit 101 determines whether or not a search condition has been received from the client terminal 402 (step S2101). Here, the reception unit 101 waits for reception of a search condition (step S2101: No).

When the search condition is received (step S2101: Yes), the specifying unit 1301 determines the similarity R _i (k) between the attribute information of the attribute Ak specified by the search condition and the attribute information of the attribute Ak of the data Di. A similarity specifying process for specifying is executed (step S2102).

  Next, the first calculation unit 1302 multi-vectorizes the search condition and the data Di using the attributes A1 to AK as vector elements, thereby calculating the intervector distance Li between the search condition and the data Di. Calculation processing is executed (step S2103).

  Thereafter, the search unit 103 executes a search process for searching for data from the item master 403 (step S2104), and the output unit 104 outputs a search result list to the client terminal 402 (step S2105). A series of processing by this flowchart is complete | finished.

  As a result, it is possible to search for data Di that may be helpful to the user even if the search condition specified by the user does not completely match the search data specified by the user.

<Similarity identification processing procedure>
Next, a specific processing procedure of the similarity specifying process in step S2102 in FIG. 21 will be described. FIG. 22 is a flowchart illustrating an example of a specific processing procedure of the similarity specifying process in step S2102.

  In the flowchart of FIG. 22, first, the specification unit 1301 specifies the attributes A1 to AK specified by the search condition received in step S2101 shown in FIG. 21 (step S2201). Next, the identifying unit 1301 initializes “k” of the attribute Ak with “k = 1” (step S2202), and initializes “i” of the data Di to be searched with “i = 1” (step S2202). S2203).

Thereafter, the specifying unit 1301 selects data Di from the item master 403 (step S2204). Then, the identifying unit 1301 refers to the similarity definition table 1400 to identify the similarity R _i (k) between the attribute information of the attribute Ak specified by the search condition and the attribute information of the attribute Ak of the data Di. (Step S2205). Note that the identified result is stored in, for example, the similarity table 1500 shown in FIG.

  Next, the identification unit 1301 increments “i” of the data Di (step S2206), and determines whether “i” is greater than “n” (step S2207). If “i” is equal to or less than “n” (step S2207: NO), the process returns to step S2204.

  On the other hand, if “i” is greater than “n” (step S2207: Yes), the identifying unit 1301 increments “k” of the attribute Ak (step S2208), and whether “k” is greater than “K”. Is determined (step S2209).

  If “k” is equal to or less than “K” (step S2209: NO), the process returns to step S2204. On the other hand, when “k” is larger than “K” (step S2209: YES), the process proceeds to step S2103 shown in FIG.

Thereby, the similarity R _i (k) with the search condition can be specified for each attribute Ak of each data Di to be searched. In other words, by performing the process described later using the similarity, even if the data does not match the search condition, it is possible to search for data that is highly likely to be useful to the user by inferring from the search condition. .

<Inter-vector distance calculation processing procedure>
Next, a specific processing procedure of the inter-vector distance calculation processing in step S2103 in FIG. 21 will be described. FIG. 23 is a flowchart illustrating an example of a specific processing procedure of the inter-vector distance calculation processing in step S2103.

  In the flowchart of FIG. 23, first, the first calculation unit 1302 refers to the axis definition table 1700 to identify the axes corresponding to the attributes A1 to AK specified by the search condition (step S2301).

  Then, in the orthogonal coordinate system in which the axes corresponding to the specified attributes A1 to AK are radially extended from the origin by the first calculation unit 1302, a reference vector representing the search condition received in step S2101 shown in FIG. An SV is created (step S2302).

  Next, the first calculation unit 1302 initializes “i” of the data Di to be searched with “i = 1” (step S2303), and selects the data Di from the item master 403 (step S2304). ). Thereafter, the first calculation unit 1302 creates the comparison vector Vi representing the data Di by referring to the similarity table 1500 (step S2305).

  Then, the first calculation unit 1302 calculates the inter-vector distance Li between the created reference vector SV and the comparison vector Vi using the above equation (3) (step S2306). Next, the first calculator 1302 increments “i” of the data Di (step S2307), and determines whether “i” is greater than “n” (step S2308).

  If “i” is equal to or less than “n” (step S2308: NO), the process returns to step S2304. On the other hand, when “i” is larger than “n” (step S2308: Yes), the process proceeds to step S2104 shown in FIG.

Thereby, based on the similarity R _i (k) for each attribute Di of each data Di, the search conditions and each data Di are converted into multidimensional vectors using the attributes A1 to AK as vector elements. The inter-vector distance Li can be calculated. That is, even if the data has attribute information different from the attribute information specified by the user, the data that is likely to be useful to the user has a short inter-vector distance. Therefore, the search device disclosed in the present embodiment can more accurately identify data that is more likely to be useful to the user among a plurality of data by using the inter-vector distance.

<Search processing procedure>
Next, a specific processing procedure of the search processing in step S2104 in FIG. 21 will be described. Here, a case where m pieces of data matching the search condition are searched from the item master 403 will be described as an example (n> m). FIG. 24 is a flowchart illustrating an example of a specific processing procedure of the search processing in step S2104.

  In the flowchart of FIG. 24, first, the search unit 103 initializes “j” representing the number of searched data with “j = 1” (step S2401). Then, the search unit 103 searches the item master 403 for data Di having the smallest inter-vector distance Li calculated in step S2306 in FIG. 23 (step S2402).

  Thereafter, the search unit 103 registers the searched data Di in the search result list (step S2403). A specific example of the search result list will be described later with reference to FIG. Next, the search unit 103 increments “j” representing the number of searched data (step S2404), and determines whether “j” is greater than “m” (step S2405).

  Here, when “j” is equal to or less than “m” (step S2405: No), the process returns to step S2402, and the search unit 103 obtains the data Di with the smallest unsearched inter-vector distance Li from the item master 403. Search is performed (step S2402). On the other hand, when “j” is larger than “m” (step S2405: YES), the process proceeds to step S2105 shown in FIG.

  Thereby, it is possible to obtain the top m search results having a high similarity to the search condition designated by the user from the data groups D1 to Dn to be searched.

  FIG. 25 is an explanatory diagram of a specific example of a search result list. 25, the search result list 2500 stores the data Di searched in step S2402 shown in FIG.

  Specifically, the search result list 2500 stores item information 600-1, 600-3, 600-4 in the item master 403 as search results. When the search result list 2500 is output from the search device 401 to the client terminal 402, for example, the search result screen 2000 shown in FIG. 20 is displayed on the display 509 of the client terminal 402.

As described above, according to the search device 401 according to the second embodiment, the similarity R _i (k) for each attribute Ak of each data Di can be specified by referring to the similarity definition table 1400. it can. Further, according to the search device 401, based on the similarity R _i (k) for each attribute Di of each data Di, the search conditions and each data Di are converted into multidimensional vectors using the attributes A1 to AK as vector elements, and the search is performed. The inter-vector distance Li between the condition and each data Di can be calculated. Then, according to the search device 401, the data Di can be searched from the data groups D1 to Dn based on the calculated inter-vector distance Li of each data Di. Thereby, even if the search condition specified by the user does not completely match, it is possible to search for data Di that may be helpful to the user.

In addition, according to the search device 401, the factor Ri (k) is obtained by a creator who is in a similar situation by obtaining a similarity R _i (k) by converting each factor related to human decision making into a multidimensional vector as a vector element. Documents and drawings can be provided to the user. Specifically, for example, it is possible to collectively provide a user with documents and drawings created by a creator who is similar in business content and problem to be faced and who has a similar method of dealing with the business and the problem. Thereby, the user can refer to the coping method with respect to the problem that other people who have similar business contents faced, and the efficiency of the business can be improved.

  Further, according to the search device 401, when there are a plurality of data D [t] with the minimum distance between vectors, the axis representing the specific attribute AP is further vectorized to identify each minimum data D [t]. The inter-vector distance L [t] with the axis representing the attribute AP can be calculated. Then, according to the search device 401, the data D [t] is searched from the data D [1] to D [T] based on the calculated inter-vector distance L [t] of each data D [t]. can do. As a result, even when there are a plurality of pieces of data having the minimum distance between vectors, the data Di to be searched can be uniquely specified.

Further, according to the search device 401, the data Di can be searched from the data D1 to Dn based on the accumulated similarity R _i obtained by accumulating the similarity R _i (k) for each attribute Ak of each data Di. it can. Thereby, even if the search condition specified by the user does not completely match, it is possible to search for data Di that may be helpful to the user.

  Note that the search method described in this embodiment can be realized by executing a program prepared in advance on a computer such as a personal computer or a workstation. The search program is recorded on a computer-readable recording medium such as a hard disk, a flexible disk, a CD-ROM, an MO, and a DVD, and is executed by being read from the recording medium by the computer. The search program may be distributed via a network such as the Internet.

  Further, the search devices 100 and 401 described in the present embodiment include a special-purpose IC (hereinafter simply referred to as “ASIC”) such as a standard cell or a structured ASIC (Application Specific Integrated Circuit), or a PLD (such as an FPGA). It can also be realized by Programmable Logic Device). Specifically, for example, the search devices 100 and 401 can be manufactured by defining the functions of the search devices 100 and 401 by HDL description, logically synthesizing the HDL description and giving it to the ASIC or PLD.

  The following additional notes are disclosed with respect to the embodiment described above.

(Supplementary Note 1) A reception unit that receives a search condition that specifies attribute information of an arbitrary attribute;
A creation unit for creating another search condition for designating the other attribute information by referring to a table defining other attribute information related to the attribute information designated by the search condition accepted by the accepting unit;
A search unit that searches for data from a data group to be searched based on other search conditions created by the creation unit;
An output unit for outputting a search result searched by the search unit;
A search device comprising:

(Supplementary note 2) The search device according to supplementary note 1, wherein the retrieval unit further retrieves data from the data group based on a retrieval condition received by the reception unit.

(Additional remark 3) Each data of the said data group has the attribute information for every attribute contained in the some attribute which characterizes the said each data,
The accepting unit accepts a search condition for designating attribute information for each attribute,
For each of the attributes, the creation unit refers to a table that defines other attribute information having the same attribute as the attribute information of the attribute specified by the search condition received by the receiving unit. Create other search conditions that specify attribute information,
The search device according to appendix 1 or 2, wherein the search unit searches for data from the data group based on other search conditions created for each of the attributes by the creation unit.

(Additional remark 4) By referring to the table which prescribes | regulates the similarity between the attribute information which has the said attribute for every said attribute, the attribute information of the said attribute designated by the said search conditions, and the attribute of the said attribute of each said data It has a specific part that identifies the degree of similarity to information,
The search device according to appendix 3, wherein the search unit searches for data from the data group based on a similarity for each attribute of the data specified by the specifying unit.

(Additional remark 5) Based on the similarity for each said attribute of each said data, the said search conditions and each said data are obtained by making said search conditions and each said data into multidimensional vector by making said some attribute into a vector element. A calculation unit for calculating a distance between vectors of
The search device according to appendix 4, wherein the search unit searches for data from the data group based on the calculation result calculated by the calculation unit.

(Additional remark 6) The said search part searches the data with the shortest distance between vectors calculated by the said calculation part from the said data group, The search apparatus of Additional remark 5 characterized by the above-mentioned.

(Supplementary note 7) When there are a plurality of pieces of data having a minimum inter-vector distance calculated by the calculation unit (hereinafter referred to as “first calculation unit”), an axis representing a specific attribute among the plurality of attributes A second calculating unit that calculates a vector-to-vector distance between each of the minimum data and the axis by vectorizing
The search device according to appendix 6, wherein the search unit searches for data having a minimum inter-vector distance calculated by the second calculation unit from the plurality of the minimum data.

(Supplementary Note 8) A third calculation unit that calculates a value obtained by accumulating the degree of similarity for each attribute of the data for each of the data,
The search device according to appendix 4, wherein the search unit searches the data group for data having a maximum value calculated by the third calculation unit.

(Appendix 9)
Accepting means for accepting a search condition for specifying attribute information of an arbitrary attribute;
Creating means for creating another search condition for designating the other attribute information by referring to a table defining other attribute information related to the attribute information designated by the search condition accepted by the accepting means;
Search means for searching for data from a data group to be searched based on other search conditions created by the creation means;
Output means for outputting a search result searched by the search means;
A search program characterized by functioning as

(Additional remark 10) The reception process which receives the search conditions which designate the attribute information of arbitrary attributes,
A creation step for creating another search condition for specifying the other attribute information by referring to a table that defines other attribute information related to the attribute information specified by the search condition received by the reception step;
A search step for searching for data from a data group to be searched based on other search conditions created by the creation step;
An output step of outputting a search result searched by the search step;
A search method characterized in that a computer is executed.

DESCRIPTION OF SYMBOLS 100,401 Search apparatus 101 Reception part 102 Creation part 103 Search part 104 Output part 110 Attribute table 402 Client terminal 403 Item master 404 Document master 800 User classification table 900 Creation reason classification table 1000 Creation method classification table 1301 Identification part 1302 1st part Calculation unit 1303 Second calculation unit 1400 Similarity definition table 1500 Similarity table 1700 Axis definition table

Claims (7)

A storage unit that stores the data group having attribute information for each attribute included in a plurality of attributes that characterize each data of the data group to be searched;
A reception unit for receiving a search condition for designating attribute information for each attribute;
For each attribute, refer to a table that defines the similarity between the attribute information of the attribute, the similarity between the attribute information of the attribute specified in the search condition and the attribute information of the attribute of each data, respectively A specific part to identify;
In the coordinate system in which each of the plurality of attributes is an axis of each dimension, a vector having the similarity for each attribute of the data specified by the specifying unit as a component of the axis corresponding to the attribute, A calculation unit that calculates an inter-vector distance with a reference vector that is a maximum value that can be taken as the similarity of an axis component corresponding to each of the attributes of
A search unit for searching for data from the data group based on the calculation result calculated by the calculation unit;
An output unit for outputting a search result searched by the search unit;
A search device comprising: The search unit
The retrieval apparatus according to claim 1, further comprising: retrieving data from the data group based on a retrieval condition received by the reception unit. A creation unit that creates another search condition that specifies the other attribute information by referring to a table that defines other attribute information related to the attribute information specified by the search condition received by the reception unit;
The search unit
The search device according to claim 1, wherein data is searched from a data group to be searched based on another search condition created by the creation unit. The creating unit
For each attribute, refer to a table that defines other attribute information having the same attribute as the attribute information of the attribute specified by the search condition received by the receiving unit, and specify other attribute information. Create each search condition,
The search unit
The search device according to claim 3, wherein data is searched from the data group based on other search conditions created for each of the attributes by the creation unit. The calculation unit includes:
When there are a plurality of pieces of data having a minimum inter-vector distance with the calculated reference vector, for each piece of data with a minimum inter-vector distance with the reference vector, the similarity for each attribute of the data is determined for the attribute. An axis corresponding to an attribute other than the specific attribute , and a vector that is a component of the axis corresponding to, and an axis component corresponding to a specific attribute among the plurality of attributes as a maximum value that can be taken as similarity Calculate the distance between the vector and the axis vector with zero component,
The search unit
The data having the minimum inter-vector distance with the axis vector calculated by the calculation unit is searched from a plurality of data having the minimum inter-vector distance with the reference vector . 5. The search device according to any one of 4. A computer that can access the data group having attribute information for each attribute included in the plurality of attributes that characterize each data of the data group to be searched,
Accepting means for accepting a search condition for designating attribute information for each attribute;
For each attribute, refer to a table that defines the similarity between the attribute information of the attribute, the similarity between the attribute information of the attribute specified in the search condition and the attribute information of the attribute of each data, respectively Identification means to identify,
In a coordinate system in which each of the plurality of attributes is an axis of each dimension, a vector having a similarity for each attribute of each of the data specified by the specifying means as a component of an axis corresponding to the attribute, Calculating means for calculating an inter-vector distance from a reference vector having a maximum value that can be taken as a component of an axis corresponding to each of the attributes ;
Search means for searching for data from the data group based on the calculation result calculated by the calculation means;
Output means for outputting a search result searched by the search means;
A search program characterized by functioning as A computer accessible to the data group having attribute information for each attribute included in a plurality of attributes that characterize each data of the data group to be searched,
Accepting a search condition for designating attribute information for each attribute;
For each attribute, refer to a table that defines the similarity between the attribute information of the attribute, the similarity between the attribute information of the attribute specified in the search condition and the attribute information of the attribute of each data, respectively A specific process to identify;
In the coordinate system in which each of the plurality of attributes is an axis of each dimension, a vector having the similarity for each attribute of the data specified by the specifying step as a component of an axis corresponding to the attribute, A calculation step of calculating an inter-vector distance from a reference vector having a maximum value that can be taken as a similarity of an axis component corresponding to each of the attributes ;
A search step of searching for data from the data group based on the calculation result calculated by the calculation step;
An output step of outputting a search result searched by the search step;
A search method characterized by executing

Images