WO2005122014A2 - 情報空間処理装置、プログラム、および、方法 - Google Patents
情報空間処理装置、プログラム、および、方法 Download PDFInfo
- Publication number
- WO2005122014A2 WO2005122014A2 PCT/JP2005/010856 JP2005010856W WO2005122014A2 WO 2005122014 A2 WO2005122014 A2 WO 2005122014A2 JP 2005010856 W JP2005010856 W JP 2005010856W WO 2005122014 A2 WO2005122014 A2 WO 2005122014A2
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- space
- node
- information
- unit
- calculation
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
Definitions
- the present invention relates to management and retrieval of an information space that is a knowledge information capability in a computer program.
- the present invention relates to an information space processing apparatus for performing analysis, sharing, and the like.
- Non-Patent Document 1 Non-Patent Document 2
- Non-Patent Document 3 Non-Patent Document 4
- Non-Patent Document 5 Non-Patent Document 9.
- next-generation information management of information such as (codename) WinFS of Microsoft's next-generation file system described in Non-patent Document 6, Non-patent Document 7, and Non-patent Document 8 is also available.
- Research and development is progressing as a method. The idea is to make the file system accessible like accessing the file system on the Web. This means that all meta information about the file is stored in a relational database, enabling advanced query search, cross-search, search speed improvement, search by classification, and narrow search of search results. is there. However, this is also a method of managing and searching for meta information of information.
- Non-patent Document 1 Jack Park, edited by Sam Hunting, edited by Tsuyoshi Goto, "XML Topic Maps—The Era is the Semantic Web"
- Non-Patent Document 2 Members of the TopicMaps.Org Authoring Group, XML Topic Maps (XTM) 1.0, [online], [Searched on May 20, 2004], Internet: http://www.topicmaps.org/xtm / 1.0 />
- Non-patent document 3 ISO 13250
- Non-patent Document 4 W3C, Resource Description Framework (RDF), [online], [May 2004 search], Internet URL: http: ⁇ www.w3.org / RDF />
- Non-Patent Document 5 W3C, OWL Web Ontology Language XML PresentationSyntax, [on line], [Searched on May 20, 2004], Internet http: ⁇ www.w3.org / TR / owHonlsynt ax />
- Non-patent document 6 MicrosoftCorporation, WinFS, [online], [May 2004 search], Internet URL: nttp: / 1 msdn.microsoft.com/Longhorn/ understanding / pnlars / Win Fb / default.aspx
- Non-Patent Document 7 By Microsoft Corporation, WinFS in the Longhorn SDK, [online], [Searched May 2004], Internet URL: http://longhorn.msdn.microsoft.com/?//lon ghorn.msdn.microsoft.com/lhsdk/winfs/daovrWelcomeToWinFS.aspx>
- Non-Patent Document 9 Computational Intelligence Part4 Knowledge Engineering Ontology Promotes Knowledge Sharing ”, says Keio Byte, February 2004, Keitsu BP, 2004, P. 42-45
- an object of the present invention is to provide an information space processing device or the like that can treat a meaning as it is, by treating the meaning of a person as state information in a space, and can treat the meaning as it is.
- an information space evaluation device is an information space evaluation device that evaluates a plurality of spaces formed by a set of nodes as a set of information.
- a data acquisition unit that acquires information from the space
- an overlay calculation processing unit that performs overlay calculation that is a process of calculating the relevance of different spaces based on the acquired information
- the overlay calculation unit A data output unit for outputting a result of the superposition calculation by the processing unit.
- meaning is treated as space.
- the meaning is treated as positional information that exists in a space that is different from some defined attribute information, and state information of its existence.
- this space is particularly called a semantic space. Also, treating it as state information of its existence plays a significant role in handling meaningful ambiguity.
- the concept of information can be efficiently handled by handling the state of existence. This principle allows for more efficient management and acquisition of information. But these To be effective, certain management and processing methods must be used.
- Spatial superposition processing is a very important concept in the present invention. This is to calculate by comparing existence in different spaces. Of particular importance is the comparison of whether this comparison is the same.
- Each space expresses a certain concept.
- the existence of the same thing in each different space means that the same thing exists in different concepts.
- the existence is understood as meaning. This existence does not depend on the structure of the space itself or the relationship between the information and the information, so that a large degree of freedom can be given to the description of the space.
- objects that exist across multiple spaces can be revealed.
- the superimposition process is a process of exposing information by comparing data bias and characteristics in space.
- superposition processing can produce a great effect by utilizing spatial predictability.
- Spatial predictability means that humans can predict unrecognized parts by recognizing information linked to the space or a part of the space. By performing the superposition process based on this prediction, it is possible to reach easily by revealing concepts or information that are not recognized, that is, unknown even if there is no clear existence. In this way, it can be said that the superimposition process is a process capable of retrieving information by utilizing human predictive ability.
- the recognition process the non-recognition process, and the superimposition process will be described with reference to the drawings.
- the perceived information power like 111 can be predicted to some extent.
- the cognitive part power also predicts the non-cognitive part, and the information about the non-cognitive part can be revealed by the superimposition process.
- Figure 5 shows the flow of information by searching using keywords and natural language processing.
- the person 121 converts his or her own image into a keyword or natural language at 122. This information is processed by the computer at 123 and output as a result 124 for human perception.
- Figure 6 shows the information flow of the overlay process.
- the human 121 maps the power of the image that he or she has in the defined space at 125, performs the computer processing on the result of the mapping at 126, and outputs the result as the result 127 to output the human. Recognizes it.
- Projection of a space is a mechanism that is useful when expressing different things while maintaining a uniform structure. In other words, it is possible to treat multiple versions of the space as one concept. Projection of space can make uniform data easier to handle. As a use, projection, which manages space for multiple languages such as Japanese, English, and Spanish, is powerful. By projecting, it is possible to clearly create a cohesive space for various aspects while maintaining a specific structure.
- the layer can provide additional information or a process using the additional information for the space.
- information about security is usually different from the concept you want to represent in space. Specifically, it is assumed that there is a space expressing a certain test result. Information on the security of this space, such as access permission and denial, should be related to the test score, V, and the test result.
- the expansion of the space is facilitated by the automatic space classification process. This process eliminates the need for humans to manually consider where a node should be added in a space and when it should exist. Can be significantly reduced.
- the present invention provides an information space evaluation method which can be realized as an information space evaluation device as described above, a program expressing the method as a step, and a computer readable CD-ROM or the like recording the program. It can also be realized as a medium.
- the invention's effect can be realized as an information space evaluation device as described above, a program expressing the method as a step, and a computer readable CD-ROM or the like recording the program. It can also be realized as a medium.
- FIG. 5 Diagram showing information flow in information search by keyword or natural language
- FIG. 6 Diagram showing information flow in information search by overlay processing
- FIG. 27 Diagram for explaining node implementation using XML
- Figure 28 Diagram showing an example of the entity of a node expressed in XML
- FIG.62 Diagram showing an example of logical operation by spatial XNOR
- ⁇ 67 Diagram showing the result of weighted search adjusted by priority weighting.
- ⁇ 68 Diagram showing an example of adjusting priority by link distance.
- FIG. 81 is a diagram showing an example of a logical operation in a data scanning process
- FIG. 82 is a diagram showing an example of weighting in data scanning processing
- FIG.91 Node diagram representing music information and distinguished by synchronization point equivalence evaluation.
- FIG.92 Node diagram representing music information and differentiated by equivalence evaluation by genre.
- FIG. 161 Implementation diagram of client implementing invention by connection of external force
- FIG. 101 is a functional block diagram showing a configuration example of the information space processing apparatus 4000 according to the present embodiment.
- This information space processing device 4000 evaluates (analyzes, manages, searches, shares, etc.) multiple spaces formed by a set of nodes, which are a set of information.
- the device is characterized by the fact that it can treat the meaning of a human as ambiguous by treating it as state information in space.
- the data acquisition unit 4210, processing start unit 4212, projection unit 4411, It comprises a calculation processing unit 4211, a data output unit 4213, and an automatic classification processing unit 4809.
- the data acquisition unit 4210 further includes a space selection unit 4250 including a link selection unit 4251 and a hierarchy selection unit 4252, and a spatial filtering unit 4253.
- the superposition calculation processing unit 4211 further includes an operation processing unit 4216 including a logical operation calculation unit 4220, a weight calculation unit 4221, an integrated calculation operation unit 4222, a filtering unit 4223, and a Karoe processing unit 4224, and an equivalence evaluation unit. It consists of 4215 and a comparative study section 4217.
- FIG. 2 also shows a spatial data section 4201 and the like existing outside the information spatial processing device 4000.
- Each component is implemented by one or more programs and Z or hardware. Further, the program may be processed by a plurality of computers.
- the overall processing flow uses the spatial data section 4201 to use the processing start section 4212 and the data acquisition section 4210 to determine the processing to be performed and the target space. Further, using the determinations of the processing start unit 4212 and the data acquisition unit 4210, the overlay calculation processing unit 4211 performs overlay processing, and the data output unit 4213 outputs the result to the output destination 4214.
- the space data section 4201 is a portion where information related to space is present, and means a computer-readable recording medium such as a memory or a magnetic disk, or a network itself. In order to access this information, it is common to read the information from the realized space on the memory or magnetic disk, or to obtain the other computer space information on the network.
- the processing start unit 4212 determines what kind of processing is to be performed. Here, it may be an instruction from an external or internal program, or a case where a human instructs a process. In this part, it is decided which space is to be processed and how to perform the overlay processing. In performing the processing, the spatial data section 4201 is accessed, and information is collected using the data acquisition section 4210 and the like. Such processing is often performed, for example, when a human interactively determines a search target through a program.
- the data acquisition unit 4210 is a processing unit that acquires information from the spatial data unit 4201. This is performed in order to obtain information to be used in the start unit 4212 and to use the information in the overlay calculation processing unit 4211.
- the data acquisition unit 4210 acquires data to be used for the overlay processing using a link selection unit 4251 or a hierarchy selection unit 4252, which is a kind of the space selection unit 4250. If necessary, the result of the data acquisition unit 4210 is subjected to a spatial filtering unit 4253 or the like to select data to be used.
- the overlay calculation processing unit 4211 is a processing unit that performs overlay calculation, which is a process of calculating the association between different spaces, based on the information acquired by the data acquisition unit 4210. Based on this information, various operations are performed using the arithmetic processing unit 4216 and the equivalence evaluation unit 4215 to obtain the truth value of the logical operation and the weight (score) of the weighting process and the weighted integrated calculation. The calculation of the truth value by the logical operation is implemented in the logical operation calculation unit 4220.
- the weight calculator 4221 obtains the weight (score) of the weight.
- the integrated calculation operation unit 4222 performs a process for weighted integrated calculation using the result of the weight calculation unit 4221. It is also effective to apply the filtering unit 4223 to the result of the logical operation calculation unit 4220, the weighting calculation unit 4221, and the result of the integrated calculation operation unit 4222, and to apply the processing unit 4224.
- the equivalence evaluation unit 4215 will be described.
- the equivalence evaluation unit 4215 is a processing unit that evaluates whether or not nodes can be regarded as equivalent.
- any equivalence evaluation that basically cannot be used such as synchronization point equivalency evaluation, equivalence evaluation using node properties and node type can be used.
- These can be defined in the scripting language 4240, or the function of equivalence evaluation can be delegated to an external program 4241.
- the comparison and examination unit 4217 performs a comparison and examination process for comparing and evaluating the nodes. Usually, numerical information is obtained by comparison.
- the comparison process can be defined in the scripting / programming language 4240, or the function of the comparison process can be delegated to an external program 4241.
- the arithmetic processing unit 4216 performs a comparison calculation for different spaces based on the evaluation by the equivalence evaluation unit 4215 and the comparison and examination unit 4217, and calculates the result as the relevance. is there.
- the logical operation calculator 4221 calculates the truth value of the logical operation. Calculation is etc. This is performed using the price evaluation unit 4215 or the comparison and reviewing unit 4217. For example, for the evaluation by the equivalent evaluation unit 4215, true / false is determined by judging whether or not a certain condition is satisfied, and true / false of different spaces is logically calculated. The result is used to select the truth value by the force passed to the data output unit 4213 and the filtering unit 4223, and to be used for the calculation by the filtering unit 4223 and the processing unit 4224.
- the weight calculation section 4221 calculates the weight (score) of the weight processing.
- the calculation is performed using the equivalence evaluator 4215. For example, the number of nodes determined to be equivalent by the equivalence evaluation unit 4215 is counted, and a value depending on the counted number is weighted as a weight for a space or a node. Normally, uniqueness processing, priority adjustment, and various calculations are performed in this part.
- the integrated calculation operation unit 4222 calculates the weight (score) of the weighted integrated calculation.
- the calculation is performed using the equivalence evaluator 4215. Usually, this calculation is performed using the result of the weight calculation unit 4221. For example, a comparison calculation for different spaces is performed using the weights obtained by the weight calculation unit 4221.
- the calculation results of the weighting calculation unit 4221 and the integrated calculation calculation unit 4222 may be output to the data output unit 4213 as they are. It is also effective to output the adjusted score using the unit 4224.
- the filter unit 4223 performs a filtering process, which is a mechanism for narrowing down results such as a logical operation, a weighting process, or a weighted integrated calculation.
- Information can be narrowed down by applying to the results of the logical operation calculation unit 4220, the weight calculation unit 4221, and the integrated calculation operation unit 4222.
- the result of the logical operation calculation unit 4220 is used, or a node judgment formula is used. It is also effective to carry out the result further in the processing section 4224.
- These results are output to the output destination 4214 by the data output unit 4213.
- the processing unit 4224 performs a weighting process or a weighting process for performing a process on the weight of the integrated weighting calculation. In general, processing is performed using the result of the logical operation calculation unit 4220 or performing processing using a node determination formula. The results can be further filtered. It is also effective to sort by the filtering unit 4223. These results are output to the output destination 4214 by the data output unit 4213.
- the data output unit 4213 receives the result of the superposition calculation process 4211 and outputs the result to the output destination 4214.
- the output destination is a computer-readable recording medium or a network. Normally, the results are stored in memory or magnetic disk, the results are transferred to another computer or program using a network, or the output is made using a display or speakers. Further, the output result can be further arranged by using a filtering unit 4223 and a cache processing unit 4224.
- the projection unit 4411 is a processing unit that enables a plurality of versions of the space to be treated as one concept, that is, a processing unit that projects the space. By using this function, information can be prevented from being dispersed. Projectionization is effective for multilingual space design and version control of complex programs and documents.
- FIG. 161 As an operation mode of the present invention, a mode in which one or a plurality of information receiving apparatuses are connected to the information providing apparatus 6005 to receive information as shown in FIG. 161 is effective.
- information receiving apparatuses 6001 and 6002 obtain information from information spatial processing apparatus 4200 and 6005 including spatial data section 4201 via information connecting sections 6011 and 6012. Accordingly, even if the client does not actually mount the information space processing device 4200, the information receiving device can use information such as the result of the superposition process.
- the implementation shown in FIG. 161 allows information to be easily shared by many users.
- An operation mode in which a part of the functions of the spatial data unit 4201 and the information spatial processing device 4000 is shared by a plurality of devices is also effective. In this way, by dispersing the functions, it is possible to obtain an advantage that the cost of calculation processing is not excessively concentrated on a specific program or computer.
- the overlay calculation processing unit 4211 is also mounted on the devices 6021 and 6022. In this case, the information connection units 6031 and 6032 acquire information necessary for the overlay processing via 6025, and the overlay processing is performed by the respective devices 6021 and 6022.
- the “node” is realized by a computer-readable recording medium or an information communication network such as the Internet.
- a node is basically a unit for linking and managing information, or information itself. Usually, information is tied to nodes and managed.
- the automatic classification processing unit 4809 is a control unit that controls to automatically execute each component included in the information space processing device 4000 sequentially.
- the "user” includes not only a person who uses the program but also a program which uses the program.
- a "link” is a mechanism for connecting nodes. It can be expressed in any form as long as a human, a program, or a device can recognize the connection between nodes.
- a link exists to indicate a relationship such as a structure of a node group.
- the link includes a unidirectional link and a bidirectional link. Although it depends on the environment in which it is used, the general implementation is easy to use because it is possible to follow the counter in both directions with a bidirectional link.
- nodes and spaces having various characteristics can be created.
- nodes can be localized in the link space.
- a typical example is a hierarchical space in which link relationships are linked by parent-child relationships.
- the real space includes a "link space” formed by nodes and a link between the nodes, and an "intangible space” formed by nodes in each category (that is, a simple set of nodes).
- link space can structure and systematize the space, it is excellent in visibility and understandability.
- intangible space can be replaced by the link space.
- implementation in intangible space is simpler and easier to implement than in link space. It is also excellent when used for temporary calculation processing, because it is not necessary to consider link-related processing, especially in terms of processing speed.
- a program that uses a space is called a “space program” or a “space management program”.
- FIG. 7 is a node diagram of the link space.
- 210 is a node and 211 is a link.
- the space formed by the links between the nodes is the link space.
- nodes are stored in an unlinked state.
- 220 is an intangible space and 221 is a node.
- An intangible space is formed by such a simple set of data nodes.
- space is a very broad term.
- a space can be called a space even with one node.
- a plurality of spaces can be called one space.
- a space is a collection of one or more nodes regardless of their structure, shape, or state.
- the space will be described with reference to FIG. 231, 232, and 233 in Fig. 9 can all be called spaces.
- 231 may be called a node.
- FIG. 10 is a diagram in a case where link information of a node is stored in the node itself.
- the node 300 holds link information to the node 301.
- 300 and 301 may refer to nodes depending on the implementation.
- FIG. 11 is a diagram showing a case where link information is described outside. Node 310 and node
- 311 is in a linked state.
- the link relationship is defined by 312 instead of the link information being stored in each of the nodes 310 and 311. 310, 3
- 11 may be a reference to a node, depending on the implementation.
- a link of a node When a link of a node is implemented on a memory, a method of expressing the link by holding a pointer to the node or a reference to an object is generally used.
- the link of the node can also be expressed in a format such as HTML or XML. It can also be expressed using HTML hyperlinks or XML Xlinks.
- “property” means information linked to a node.
- Information linked to a node can be called a property. What is set in the property is left to the implementer of the program. For example, text such as name and description, numeric data such as ID, application-specific data format, binary data, specific memory area, XML-created schema, file list, and file stream may be used.
- properties to information of other programs and information on the network, information management in a distributed environment can be easily realized.
- synchronization is regarded as the same, but is different in space.
- node A1 and node A2 are synchronized with each other. Except for the link, the one that applies to A2 even if it is applied to A1 is called synchronized. Synchronization only needs to be able to treat those synchronized with each other as the same except for the link.
- nodes that can be regarded as the same, but differ in space, and have the property that changes in one are adapted to the other are synchronized with each other!
- synchronization points refer to things that are synchronized with each other.
- the synchronization point will be described below with reference to FIG. Node A is synchronized with each other at 400, 401, 402.
- 400, 401, and 402 are called the synchronization points of A, and the synchronization points are said to be the same.
- the synchronization point can be regarded as asynchronous. However, in actual implementations, it is often better to treat synchronous and asynchronous as clearly separated.
- a reference generally exists as a pointer to an entity on a memory, and exists as an object or an object reference in an object-oriented environment. Link information defined outside can also be referred to as reference.
- To multiplex operations is to express synchronization by simultaneously applying one instruction to a plurality of synchronized nodes.
- FIG. 13 illustrates a reference management type implementation of an instance. These 421, 422 are references to entity 420. Therefore, the effects of operation 423 and operation 424 are both operations on the entity and the synchronization is maintained.
- FIG. 14 shows an implementation by multiplexing operations.
- 410 and 411 are different entities, but one operation 412 is duplicated and applied to multiple entities to synchronize. Has been realized.
- the method of associating information with a node or a space depends on the implementation of the space. There are two main ways of thinking when linking information to node-space. This is ⁇ synchronous information addition '', which is the addition of information to the information including the synchronization of the nodes, or not including the synchronization! /, And ⁇ asynchronous information addition '', which is the addition of the information! That's it. Synchronous information addition is information that is shared between synchronized ones, while asynchronous type information addition means that information is not shared between synchronized ones. These mechanisms are important concepts that also affect projection, layering, and automatic classification. In the present invention, a property can be said to be information to which synchronous information is added.
- FIG. 15 shows a method of linking information in an implementation that implements synchronization by reference management.
- 440 means the entity of the node, and 441 and 442 mean the reference of the entity.
- Attached calorie 443, 444, and 445 are ⁇ , 440, 441, and 442 inside or outside ⁇ ⁇ .
- Synchronization achieves synchronization by making the operation by reference an operation to the entity. Operations on references apply to 440, except for links.
- an implementation that links to the entity 440 like the additional information 445 is generally used. Further, it can also be realized by multiplexing operations on the additional information 443 and 444.
- this is a method in which the operation on the additional information 443 is also applied to the additional information 444.
- the calorie is added to the blue report 443, 444.
- the information-bearing kama becomes more flexible. For example, if the reference is information that defines a space, the user is free to attach information to a specific space.
- FIG. 16 shows a method of linking information in an implementation that achieves synchronization by multiplexing operations. Synchronization is maintained by simultaneously applying one instruction 439 to the entities 431 and 432 of the nodes in a synchronous relationship. Attached Caro Blue Report 433, 434, 435, 436, 437 is information linked inside or outside 431, 432. In this case, when asynchronous information addition is performed, information may be added to the additional information 433 and 434. When applying the synchronous information addition, it can be used by sharing the common additional information 437 from 431 and 432. Synchronous information addition can also be realized by using additional information 433 and 434. In this case, the node Just as the operations for the additional information 433 and 434 are multiplexed, the operations can also be realized by multiplexing.
- independent space node the space itself can be treated as a node.
- independent space nodes can be treated as nodes or as spaces.
- 500 and 501 are independent space nodes that are synchronized with each other. Although 500 and 501 are linked like nodes, a space is formed inside. A change to node 502 is a change to 503. This is because the inside space is synchronized because 500 and 501 are synchronized.
- Node type means a type representing a node.
- a node that embodies the node type can be said to be a node. This is similar to the relationship between classes and instances used in object-oriented languages.
- An instance is a representation of a class.
- a node can use "inheritance" to define a new node.
- This feature extends a node so that new nodes can inherit the properties of the source.
- Inheritance is often used in fields other than programming languages, which are often used in object-oriented languages. This mechanism prevents duplication of implementation, systematizes concepts, and enables expression of polymorphism in programming languages. As a concrete example, it is used for purposes such as creating a file data node that inherits nodes and stores files in nodes. This mechanism is not required for a space program, but it makes space management easier.
- Node 602 is created by inheriting node 601.
- the pronotities 615 and 616 of the node 602 inherit the properties 610 and 611 of the node 601 as they are.
- a new property 617 has been extended.
- Figures 19 and 20 are XML that defines the node type.
- FIG. 19 indicates the node 601 in FIG. 18, and
- FIG. 20 indicates the node 602 in FIG. In this case, among the properties in Fig. 20, those defined in Fig. 19 are omitted ( It is strong.
- node There are many possible application implementations of the node. By linking a URI or global ID to the property of a node and connecting it to a network, it becomes an information management system on the Internet easily. When linked to file system files and directories, it becomes a powerful file management system.
- the node may be the information itself. That is, it may be simply text data or image data. Whether these are treated as nodes can be controlled depending on the program.
- a node having a parent-child node link relationship is particularly called a "hierarchical space”. Spaces created by parent-child relationships are easier to systematize, and therefore have better learning comprehension than randomly linked spaces.
- the advantage of hierarchical space is that it is rich in “spatial predictability”. This means that it is easy to predict the space connected to it only by recognizing a part of the space. This spatial predictability has a great effect on spatial overlay processing.
- the hierarchical space is a space having a parent-child relationship
- a node serving as a starting point exists. This is called the root node.
- the target space can be specified by the path from the root node.
- Node 812 is a child node of node 811
- node 811 is a child node of node 810. What is constructed in such a parent-child relationship is a hierarchical space.
- nodes have links to information about the diary. If you try to read the diary of October 6, 2001. Do you search for 801 nodes by date or 802 memories? You will try to find nodes by date. This is space It is predictive. This means that by recognizing a part of the space, the composition of the space can be predicted.
- This information space processing device 4000 can perform a search using a property. This search can be described in a simple manner by describing recursive processing in a programming language or by using an agent that searches around.
- information between nodes can also be defined.
- a method of processing based on information given between nodes is easy to handle for a specific application, but may be difficult for other applications to handle.
- the information may be stiffened.
- the advantage of this method is that it is easy to describe automatic processing of information and inference processing by computer.
- a link node 822 is arranged between the nodes 821 and 823. This is disclosed to the actual user as a space 829. In this case, node 821 can be expressed as node 824, and node 823 can be expressed as node 826.
- the link form is also used as link node 822 to link 825.
- the information space processing apparatus 4000 is realized by a single computer or a plurality of computers.
- Figure 1 shows a schematic diagram of a general computer.
- the present invention is applicable to any computer and system, such as a personal computer, a server computer, a mainframe computer, a network system, a personal digital assistant, and an embedded microcomputer. It can be implemented in the system. Because the principle of the present invention is relatively simple, it can be implemented on a computer that can be called a stored program method, or on any computer using it.
- Object Model can also simulate the implementation of the space.
- What is called a network type database can also be said to be an implementation example of the link space.
- a hierarchical database is an implementation example of a hierarchical space.
- RDF is also an example of implementing link space.
- the file system can also be said to be a hierarchical space formed by directories and directories, and links based on parent-child relationships between directories and files.
- a node When a node is implemented on a memory by a programming language, in the case of C language, a list structure including a data part and a pointer part is generally used. In an object-oriented language such as the JAVA (registered trademark) language, it is common to create a link structure using objects.
- JAVA registered trademark
- Nodes can also be realized on a network. This means that nodes and links may be signals flowing through the network.
- FIG. 27 An implementation example when the space shown in the node diagram of Fig. 27 is created in the XML format is shown.
- the same alphabet is synchronized with each other.
- Each node is a simple node with only "name” and "type” properties.
- the definition of each node itself is shown in Figure 28 as Node A, Figure 29 as Node B, Figure 30 as Node C, Are defined in XML.
- the definition of the link when these nodes are a network type link space is the definition in Fig. 31. When these are expressed as a hierarchical space which is a kind of link space, It can also be represented as shown in Figure 32. This is a representation of the parent-child link relationship as XML parent and child elements.
- the nodes and the space need not actually exist on a computer-readable recording medium or an information communication network.
- a logical entity that can be expressed by a query, a program, a calculation formula, or the like may be called a node or a space.
- Figure 149 is just a list of information.
- the In Figure 149 only the alphabets and the order of the alphabets are displayed. However, in response to this information, "The alphabets are numbered in alphabetical order, and the numbers given are the magnitudes, k, d, s, c, f, p, x, a, e, g, m, s, Generate a space for the binary search tree by inserting elements into the river pages of v and z ", and the space will be as shown in Figure 150. In this case, however, the space does not exist as a reality until it is calculated. In this way, the calculation formula can be used in the same manner as a recording medium that can be read by a computer even if it does not actually exist, or a space that actually exists on an information communication network.
- nodes and data associated with the nodes There are countless methods for managing nodes and data associated with the nodes. All information may be stored as it is in the node, but management by the file system or serialization of data may be used. In addition, implementation by linking with a relational database is also conceivable. It is also an effective implementation to expand the spatial information on the memory when the program is executed and record it on a non-volatile recording medium for storage, so that the information is made permanent.
- the 910 node information is converted into a 912 byte stream or XML (see FIG. 34) by the 911 serialization process and stored in the 913 recording medium.
- the 913 can be a file system, a network file system, a database, or your own storage system!
- nodes An example of management of node information by the relational database by the information space processing apparatus 4000 will be described with reference to FIG. 920 and 921 are nodes.
- the properties of the node can be stored in table 922 of relational database 923 to speed up the search by properties.
- the nodes themselves can be stored in a table as data in the serialization process, making management easier. It is also useful to store the reference information of the nodes in a table.
- FIG. 36 A case will be described with reference to FIG. 36 in which spatial information is developed on a memory at the time of executing a program and is recorded on a non-volatile recording medium for storage, thereby making the information permanent.
- the space program 930 When the space program 930 is running, the space is stored on the storage medium 931 such as a faster memory. And perform calculations.
- the storage medium 931 such as a faster memory.
- efficient computer resources are effectively used.
- FIG. 15 denotes a network (information communication network) such as the Internet, an intranet, and wireless communication. 10, 11, 12, and 13 mean independent computers connected to the network.
- An address call is a method of accessing a part describing a specific process using a pointer in a memory. This is also used in C language for what is called callback processing. In most cases, methods 3, 4, and 5 are implemented using methods 1 and 2.
- 1201, 1202, and 1203 mean programs.
- the node 1210 managed by the program 1201 is used as the node 1211 in the program 1202 by the sharing means 1205.
- the node 1213 managed by the program 1203 is used as the node 1212 in the program 1202 by the sharing means 1206.
- programs 1202 can operate as if they were managed by 1202.
- Nodes 1220 and 1221 are linked by link means 1232. In this way, the space is formed by the direct link between the nodes.
- An event in space is a mechanism for transmitting a change in space, an instruction to space, and a time transition to external and internal programs.
- Figure 40 shows the appearance of the event that has occurred.
- the event that occurred in the spatial program 1301 on the computer 1300 is transmitted to the internal program as a message 1310.
- Message 1311 is passed on to external program 1302.
- Message 1312 runs through network space 1305 and is passed on to programs on other computers 1303.
- Operations 1420, 1421 represent instructions for space 1430.
- the user 1410 can execute the operations 1420 and 1421 because the user 1410 has the right to perform the operations.
- User 141 1 can perform operation 1421 but cannot perform operation 1422.
- User 1412 cannot perform operations 1420 and 1421. Thus, it imposes restrictions on the instructions that can be applied by the user who uses it.
- Link selection is a method of selecting a space from a certain node in a link space according to a specific rule. It is convenient to use the selected space for things that specify the space for processing and operations, such as spatial overlay processing, logical rules, and layers. Strictly speaking, choice is a space that is recognized as a space program. In other words, it means that the program has acquired enough information to process the space.
- Hierarchy selection is an effective selection method in a hierarchy space. This method can be used on the assumption that the link relationship is composed of a parent-child relationship. When a parent node in the hierarchical space is selected, the child nodes are also selected.
- 1520, 1521, 1522, and 1523 are nodes, and 1510, 1511, 1512, and 1513 are spaces.
- the space 1510 is selected.
- the space 1511 is selected
- the space 1512 is selected
- the space 1513 is selected.
- spatial filtering section 4253 can perform a filtering process on the selected space and extract a more necessary part of the space.
- the filtering process needs a part to determine whether it is to be selected. In this part, a node judgment formula is used. From a simple formula to a complicated node calculation formula, it can be used for filtering.
- the "node calculation formula” defines a procedure for obtaining or using information related to a node to calculate.
- the information enumerated below is representative of the information linked to the node.
- Node property information The value of a specific property held by the node, or information on linked resources
- node synchronization point information By using "node synchronization point information”, it is possible to know whether or not a particular node is a synchronization point, and how many synchronization points of a particular node are in space.
- Node property information refers to information about the properties that a node holds, the types of information that can be stored in node properties, and the definition of properties. .
- Information on node links refers to information such as how many links exist for a node and whether a node link is bidirectional.
- the "node determination formula” defines a procedure for calculating using information related to a node, such as that used in a node calculation formula, and determines whether a node requires or does not. It means the expression. Here, it may be as simple as determining whether the value of a property is a specific value or not, and using a more complicated node calculation formula to find the answer. In an actual implementation, it would be common to allow node expressions to be defined in scripting languages or programming languages. It is also effective to implement a node judgment formula by an external or internal program.
- the node calculation formula and the node determination formula used by the information space processing apparatus 4000 will be described with reference to some examples.
- the detailed definition of the node 1616 is 1601.
- An example of calculation using the property value of the node for the node 1616 is shown in the table of FIG.
- Equivalence evaluation performed by the equivalence evaluation unit 4215 will be described.
- Equivalence means the same, and equivalence evaluation means determining whether they are the same. What is equivalent changes in situations where space is used. It is common to evaluate using information related to nodes, such as those used in node formulas. There is a wide variety of methods, such as how to compare information linked to nodes as they are to determine whether they are equivalent, and how to determine more complex node formulas by defining them in a programming language or script. Listed below are those that are considered particularly important.
- FIG. 49 shows a result table that shows an example of the case where various equivalence evaluations using properties and node types are applied to these two types.
- the values of properties 1715 and 1725 can be analyzed by text mining or parsing to determine if they are equivalent.
- FIG. 52 is a result table showing an example of a case where the equivalence evaluation is applied using the information of the synchronization point and the information of the link.
- a method of performing equivalence evaluation using data outside the node is generally a method of comparing data related to the node as it is or by calculating the result.
- How to implement the equivalent mechanism is left to the implementer of the spatial program. Also, what is equivalent depends on the situation where the space is used and the design, so it is desirable to provide a mechanism according to the situation.
- the equivalence evaluation by the equivalence evaluation unit 4215 is a particularly important concept in the superposition processing and the logic rule. When information is analyzed using a spatial program, various viewpoints can be changed by changing the criteria for equivalence evaluation.
- Equivalence evaluation is relatively easy to describe externally because it only needs to pass two nodes to be compared, or node information, and pass information for determining whether they are the same or for calculation processing.
- the “overlapping process” of the space by the overlay calculation processing unit is to extract information from a plurality of spaces by using a comparative evaluation for comparing nodes. By using this processing, various information can be revealed.
- the most important thing to be used as a comparative evaluation mechanism is the equivalence evaluation by the equivalence evaluation unit 4215.
- the superimposition processing based on equivalence is also referred to as “equivalent superimposition processing”.
- the "logical operation" of a space by the logical operation calculation unit 4220 is a type of superposition processing performed using a specific node or space. In this operation, a method of expressing information of a space as a truth table for logic operation and performing a logic operation between the truth tables is general.
- a common method is to determine whether the number of nodes determined to be equivalent by the evaluation is within a specified range or not. For example, true if the equivalent number is 1 or more, false otherwise And so on. These values are better implemented by the user, depending on the situation. It can also be calculated by changing the number specified by the space.
- the score of the weighting process associated with a certain node is 1 or more, 10 or less is true, and the others are calculated as false.
- FIG. 53 A truth table in which space 1800 is replaced by a logical table is shown in FIG. 54, and a truth table in which space 1801 is replaced is shown in FIG. The details of the logical operation using the AND operation are shown in FIG.
- the calculation of the boolean value is not limited to simply expressing the space by the truth. Naturally, it is possible to apply a bit mask, perform a logical operation between the operation results, and apply a NOT operation to a specific part.
- the space weighting process by the weighting calculation unit 4221 uses the comparative evaluation of nodes. In the search, one or more spatial forces also determine the weight of the data. Equivalence evaluation is the most important thing used as a comparative evaluation mechanism. In the present invention, the weighting processing based on the equivalent evaluation is also referred to as “equivalent superposition processing”. In addition, the weighting process based on the comparison process for obtaining weights by comparing nodes is referred to as “comparison process”.
- the equivalent weighting process performed by the weighting calculation unit 4221 performs an operation by counting the number of duplicates that have been made equivalent to a specific node or space by the node equivalent evaluation.
- the overlapping number of levers is information that can be regarded as importance to the information.
- the weighting process returns information on nodes equivalent to the obtained weight (score) as shown in FIG.
- Merging is a process of merging equivalent objects that exist in space, that is, putting them together. Specifically, when there are six equivalents in a space, they are reduced to two. What is particularly important in the merging process is the unique process.
- merging is a process in which overlapping items are put together and calculated.
- Uniqueness processing is processing in which, when there is a plurality of equivalents in each space, this is regarded as a duplication number 1. For example, nodes evaluated as equivalent by the equivalence evaluator 4215 are counted as one node in each space.
- the value of the weight can be changed according to the situation. This is particularly called priority adjustment.
- This feature allows for semantic adjustments when using weighted processing. For example, adjustments to change the dependency when determining the numerical power weight of the counted nodes according to the space or the node.
- FIG. 64 This figure shows the counting of multiple numbers in three spaces.
- the nodes marked with alphabetical letters a and b be the nodes selected for search in the weighted search. It is assumed that the a nodes 1901 and 1902 in each space are equivalent, and the b nodes 1903, 1904, 1905 and 1906 are equivalent.
- the results of the weighted search for the spaces 1910, 1911, 1912, and 1913 are shown in FIG. 65 as a list of equivalent nodes and weights (scores).
- the space 1910 is calculated as the score of the a node is 1
- the space 1911 is calculated as the score of the a node is 1
- the score of the b node is 1
- the space 1912 is calculated as the score of the b node ⁇ . Therefore, the score of the superposition by adding the three spaces is 2 for the a node and 4 for the b node.
- the nodes of equivalence evaluation in space are counted as they are at the time of calculation.
- the results of the weighted search for the spaces 1910, 1911, and 1912 are shown in Fig. 66 as the list of equivalent nodes and the weight (score).
- the space 1910 is calculated with the score of the a node being 1
- the space 1911 is calculated with the score of the a node being the score of the b node
- the space 1912 is calculated with the score of the b node being 1. Therefore, the score of the equivalent weighting process by adding the three spaces is that the score of the a node is 2 and the score of the b node is ⁇ .
- the score of the b-node is merged to 1 in the space 1912.
- the weight calculation unit 4221 can use the normal weighting process and the weighting process using the uniqueness process in combination. For example, in the case of Fig. 64, in the case of weighting processing of the space 1910, the space 1911, and the space 1912, the spaces 1910 and 1911 can be calculated by weighting, and the space 1912 can be calculated by applying the uniqueness processing and weighting. is there.
- a calculation method of the unique processing will be described. There are multiple equivalent nodes in one space. In this case, if the weight is forcibly set to 1, the information on the priority may be lost. Therefore, as a method of applying the uniqueness processing, a method of performing weighting processing and then dividing by the number of equivalent nodes is preferable. With this method, priority adjustment and uniqueness processing can coexist. In actual implementation, it is better to implement both the method of forcibly setting the weight to 1 and the method of dividing the weight by the number of equivalent nodes as different methods.
- the "distance priority method” in which the priority is adjusted according to the distance from a specific node to calculate the weight may be effective.
- FIG. 68 it is assumed that the portion indicated by the broken line is the space to be subjected to the weighting process.
- the formula of "distance of 1Z link + 1" is applied as the weight of each space with reference to node 1991.
- Fig. 69 shows a method in which 0.25 is subtracted every time the link distance increases until it becomes 0. In this way, the weight is adjusted in various ways according to the link distance.
- the distance priority method it is possible to change the weight based on spatial positional information.
- Filtering that uses a node judgment formula to select necessary and unnecessary ones is an effective filtering process. This function makes it possible to perform filtering processing such as leaving only those that have a specific property. It is also effective to perform a filtering process by the filtering unit 4223 using the truth result of the logical operation.
- weight processing By using the "weight processing" by the processing unit 4224, it is also effective to change the weight for the weight processing result. This process is to pray and make changes to the results of the weight calculation. Changing the weight means adding a value to the weight or using it as an operand of a more complex operation expression to derive a new result. More specifically, a node judgment formula is used to perform processing such as adding a weight of 5 to only those nodes that meet specific conditions or those that are true using logical operations.
- the handling of the independent space node is slightly ambiguous. This is because an independent space node can be regarded as a mere node including a function as a space and a plurality of nodes. Therefore, if the independent space node is considered as a space, if it is considered as a node, it is better to implement an implementation that combines both.
- the superposition calculation processing unit 4211 performs an equivalent superposition process on the space including the independent space node.
- a weighting process is performed for the space 1985 and the space 1986 in FIG.
- the independent space nodes 1981, 1982 are synchronized with each other.
- the result of performing the equivalent weighting process on the independent space node from the viewpoint of both the node and the space is the “addition score” item in FIG.
- the part with the equivalent number A is the part processed as a node.
- step S101 the nodes in the space and the node information necessary for the superimposition processing are expanded to the data container to make the data already sorted.
- a data container is a container in the C ++ language, a collection in the Java language, and the like. Normally, the data container has a memory resource area, list structure, tree structure, hash structure, etc. inside. It is preferable that the data container be sorted after the storage by storing the data container.
- a container having a tree-structured or hash-structured data storage area inside and that can retrieve information in a sorted state when stored may be used. This is commonly called an associative container in the C ++ language.
- a calculation and a merge process are performed on the sorted data containers while scanning, and a final result is output.
- the sorted data containers it is possible to perform the calculation of step S102 in one scanning process for each data container, so that the cost of this process is reduced in linear time. It is possible to do.
- a sort criterion for performing a sort process on a data container may be a form in which adjacent ones determined to be equivalent by the equivalence evaluation used in the overlay process are adjacent.
- Figure 78 shows the implementation of the algorithm when performing the overlapping process of three spaces.
- First empty by processing 1950 The node between them or the information of the node is stored in the data container.
- Processing in data container Performs sorting in 1951.
- Processing 1950 and 1951 can be combined into one using an associative container.
- the process 1953 the data container is scanned, the weight is calculated for each space, the priority and the unique weight are calculated, and the results are combined by the merge process.
- Figure 79 shows the part of the two sorted data container powers that derives the calculation results.
- the numbers in the data container in Fig. 79 indicate nodes or node information, and those with the same number are judged to be equivalent by the equivalence evaluation.
- the data container 1 and the data container 2 are alternately scanned, and the calculation process is performed while outputting the result.
- the calculation by scanning since equivalent nodes or node information are arranged next to each other, calculation is performed while scanning, and when the non-equivalent part reaches the inside of the container, the scan of merge processing is switched. good.
- FIG. 80 shows a state where the data container is scanned.
- Those of the same alphabet mean equivalent nodes or node information.
- Figure 82 shows the calculation results when the figure is weighted using unique processing and the priority of node c is doubled.
- the processing cost is higher than that using the equivalence evaluation. It may also be effective to do Abstract data is when the node information is text-based information or multimedia. Since the weighting based on simple equivalence evaluation uses a sharp criterion of the same power, if the equivalence evaluation considers abstract contents, the processing result is unstable and what the process executor expects In some cases. Also, the nodes are organized in space V. In any case, the comparison processing is effective.
- the weight is 10 if they are the same, the weight is 8 if they are quite similar, the weight is 6 if they are similar, the weight is 4 if they are slightly similar, and the weight if they are almost similar Is 2, and if they are not quite similar, the weight is 0, and so on. It is up to the implementer to determine what weights to use during the actual comparison.
- the space superimposition process using the comparative study process is particularly called “comparison study superimposition process”.
- it is the process of performing the superposition calculation by comparing the nodes.
- the process of obtaining the weight by the weight calculation unit 4221 is referred to as “comparison weighting process”
- the process of performing a logical operation and obtaining a truth value is referred to as “comparison logical operation”.
- the weight obtained by the comparative examination weighting processing is used as a comparative examination logical operation to be used as a logical operation, or a comparative examination weighting processing to perform four arithmetic operations in units of space.
- the comparative study weighting process by the weighting calculation unit 4221 will be described.
- a typical method is to use four arithmetic operations with the weights obtained in various comparison processing as operands. In general, four arithmetic operations are performed on the basis of equivalence evaluation.
- the full comparison type covers all the space to be compared.
- the outer comparison type performs comparison only with the external space.
- the inner comparison type performs comparison only in the inner space.
- the external comparison type performs the comparison operation only with the external space. That is, the nodes included in the space 5001 are not compared with other nodes in the space 5001, but are compared only with the nodes in the space 5002. Conversely, the node in the space 5002 does not compare with other nodes in the space 5002, but compares only with the node in the space 5001.
- Figure 138 "External comparison type score”. For example, when considering node a, a comparison is made with nodes h, i, j, k, and 1 in space 5002. The specific calculation is “9 + 3 + 2 + 1 + 1” in the case of simple score addition, and the score is 16.
- the inner comparison type by the weight calculation unit 4221 will be described with reference to the drawings.
- the inner comparison type does not compare the spaces with each other, but makes a comparison between nodes inside the space. That is, the nodes included in the space 5001 are compared only with the nodes in the space 5001, and the nodes included in the space 5002 are compared only with the nodes in the space 5002.
- node a which compares with nodes a, b, c, d, e, f, and g in the same space 5001.
- the specific calculation is “3 + 2 + 2 + 4 + 1 + 3” in the case of a simple score addition, and the score is 15.
- the weighting calculation unit 4221 performs comparison using the entire space. Therefore, characteristic information is derived as a whole as a process.
- the weight calculator 4221 compares only the external space. Therefore, the external comparison type has the meaning of extracting important things in the relationship between multiple spaces. For example, as shown in FIG. 139, when comparison and weighting processing is performed in a large space 5012 and a small space 5011, the difference between the characteristics of the full comparison type and the external comparison type becomes clear. Since all comparison types tend to acquire important information as a whole, the characteristics of a small space may be buried in a huge space.
- the inner comparison type performs comparison only in the internal space, and is not affected by other spaces. It can extract the characteristic part of each space. This is a usage method used for comparison and examination logical operations, etc., rather than as a unit for comparison and examination weighting. Can be said to be normal.
- the selection / comparison type performed by the weight calculation unit 4221 will be described.
- the selection comparison type is a method of deciding whether or not to compare using a node judgment formula. In other words, the method is to calculate the comparison process when a specific condition is satisfied, and to ignore it when it is not. For example, the comparison is performed only when specific information is stored in the property of the node, and the others are ignored.
- the calculation of the comparative examination weighting process by the weighting calculation unit 4221 is not limited to the addition. Various calculations can be performed as the operands of the calculation for the weight of the comparison weighting process.
- the weight calculation section 4221 performs the calculation using the result of the comparison weighting process having a different comparison format as a calculation operand.
- the weighting calculation unit 4221 performs comparative study weighting processing.
- the result is obtained for the information of the space 5201.
- the score at the time of comparison shall conform to Figure 137.
- the calculation that means that the characteristics of the spaces 5201 and 5202 as a whole are important, and that those that are closely related to the space 5203 are particularly important, and those that are closely related to the space 5204 should be excluded. Suppose you do.
- the weight calculation unit 4221 performs a full comparison in the space 5201 and the space 5202, adds a double value in the space 5203 by the outer comparison, further performs an outer comparison in the space 5204, and subtracts the value. It can be obtained by performing the following calculation. The results of this calculation are shown in the table in FIG. FIG. 141 stores the formulas and results of the calculations actually performed. For example, in the item “Compare all 5201 and 5202”, “2 + 1 + 9 + 4 + 3” is the calculation, and “19” is the final result.
- a unique comparison process can be performed in which a node in the space is made unique by a specific equivalence evaluation before calculation.
- the most important equivalence evaluation among these is the synchronization point equivalence evaluation. Synchronization means that the information is the same entity I can say. Considering the same entity as one means eliminating or reducing duplication of information.
- the comparative study weighting processing may require complicated processing. This is because various analytical processes and data mining must be used to return reliable results. Therefore, it is desirable to implement this part so that it can be expanded by performing external force processing. For example, using a script or a programming language 4240, using an external program 4241, or using an internal program or system.
- the comparison weighting process by the weight calculation unit 4221 takes a long time for the calculation as the space size increases. Assuming that the number power of the nodes existing in one space is S "n”, the comparison and examination process based on the full comparison of the two spaces performs "O (nXn)" comparisons. That is, the order is the square of n. Furthermore, when the number of spaces is large or the cost of the comparison process itself is high, the calculation cost is further increased. For this reason, although there is a possibility that the accuracy may slightly decrease, the weighting calculation unit 4221 performs a comparative examination weighting process in which nodes that are determined to be equivalent by dividing nodes according to the criteria for equivalence evaluation are referred to as “equivalent weighting processing”. It can be said that it is meaningful to increase the calculation speed by using the narrowing comparison weighting process.
- Fig. 143 is a view of the space shown in Fig. 142 in the evaluation of equivalence based on a specific standard. Nodes that have been assigned the same alphabet are the nodes determined to be equivalent. Fig.
- the weights to be integrated can be performed between various types of weighting processing. For example, there are an equivalent weighting process and a comparative weighting process. Alternatively, the result of the weighted integration calculation may be used. In short, if weight information is linked to a node, weighted integration calculation is possible.
- the advantage of performing the weighted integration calculation is that simple weighting processing can derive powerful data, so that data for deeper analysis can be generated.
- the score of the weighting process can be regarded as a measure of importance. By adding the scores, it can be inferred whether a particular node is important for multiple equivalence evaluations or comparisons. Furthermore, by performing subtraction using a score of a specific equivalence evaluation, it is possible to give a meaning that excludes the equivalence evaluation and comparison. Furthermore, by combining complex calculations, it becomes possible to perform integrated calculations with subtle meanings.
- the procedure of the integrated calculation by the integrated calculation operation unit 4222 will be described below.
- the weights are connected to the nodes of the sky f3 ⁇ 45611, 5612, and 5613 in Fig. 145!
- the weights can be derived as a result of the superposition process or not.
- Letters that have been shaken are those that are determined to be equivalent by a certain equivalence evaluation.
- the integrated calculation by the integrated calculation operation unit 4222 can be calculated based on the criteria for equivalence evaluation.
- Integrated calculation is a process that integrates the weights obtained by multiple methods.
- integration refers to summarizing based on specific criteria.
- the criterion for integrating the weights is equivalence evaluation.
- any equivalent evaluation can be used, and deciding which of the myriad of equivalent evaluations to use can be said to be case-by-case.
- weight is the same space, it can be used by integrating weights obtained by different methods and criteria.
- the weight table may or may not be obtained by the superposition process.
- FIG. 147 those having the same alphabet are the equivalent standards to be integrated.
- the integrated calculation operation unit 4222 integrates the three, but here, it is added twice because the elements of the weight table 5623 are important.
- the calculation in this case is shown in the table in Figure 148.
- the specific calculation procedure leading to the result is shown in the “calculation” item in FIG. 148, and the specific result is stored in the “score” item. This can be expressed as x + y * 2 using a calculation formula.
- Figure 83A shows the space used for the weight integration process.
- Figure 83B, Figure 83C, Figure 83D, and Figure 83E show the space of Figure 83A with different equivalence evaluations. In these spaces, those with the same sign are equivalent.
- the table in Fig. 84 shows the results of the equivalence evaluation 1 for the space 2031 and the space 2032 in the equivalence evaluation 1
- Fig. 85 shows the results in the equivalence evaluation 2 for the space 2031 and the space 2032.
- FIG. 86 shows the result table of FIG. 86 obtained by integrating and calculating the results of FIG. 84 and FIG. 85 based on the completely independent equivalent evaluation.
- the “completely independent equivalent evaluation” is an evaluation that treats all nodes existing in the space as different from each other.
- the "equivalent node” item uses completely independent equivalence evaluation, thus treating all spatial nodes as different.
- the score is 2 because of the equivalent bl.
- the score is 1 because 2011 is equivalent b2
- the score is 2 because 2012 is equivalent a2.
- the merge score shown in FIG. 86 is obtained by simply adding the “equivalent evaluation 1 score” and the “equivalence evaluation 2 score” by the integration calculation operation unit 4222 in node units. In other words, if we explain the details of 2011, “equivalent evaluation 1 score” is 3, and “equivalent evaluation 2 score” is 1, and the sum of these is 3 + 1, and the answer is 4.
- the difference score in FIG. 86 is obtained by the integration calculation operation unit 4222 subtracting the “equivalent evaluation 2 score” from the “equivalence evaluation 1 score” in node units. Furthermore, in the “difference absolute value score”, the integration calculation operation unit 4222 calculates the absolute value of the “difference score”.
- FIG. 87 shows the result table of FIG. 87 obtained by integrating and calculating the results of FIG. 84 and FIG.
- the equivalence nodes are 2013 and 2021.
- the score of 2013 is 3 and the score of 2021 is 2, so that the score of “evaluation score of 1” is 5 by “3 + 2”.
- the score of 2013 is 1 and the score of 2021 is 3, so "1 + 3" is obtained, and the score of "equivalence evaluation 2" is 4.
- the sum of the scores of Equivalent Evaluation 1 and Equivalent Evaluation 2 is the “merge score” item.
- the "merge score” is divided by the "equivalent node” by the "score per piece".
- the weighting used to derive FIG. 86 or FIG. 87 simply obtains a score for existence, but the original weighting process is naturally performed by the weight calculation unit 4221. It may be a weighting process using a conversion process, or a process with a priority added.
- FIG. 88 shows the result of the unique weighting performed for both the spaces 2031 and 2032 by performing the unique processing.
- FIG. 89 shows the pre-processing before performing the integrated calculation, and
- FIG. 90 shows the result calculated using FIG.
- the “merge score” in Figure 90 is simply the sum of the three equivalent evaluation scores in Figure 89.
- the score of criterion 1 is operand x
- the score of criterion 2 is operand y
- the score of criterion 3 is Assuming that the operand is z, it can be expressed by the expression “x ⁇ y + z * z”.
- the filtering process by the filtering unit 4223 and the weight adding process by the processing unit 4224 can naturally be used for the integrated calculation of the weighting.
- FIG 91 is a space that expresses information about music CDs, and nodes represent music CDs. Those assigned the same number are determined to be equivalent by the synchronization point equivalent evaluation.
- Figure 93 lists the weights of each space.
- the executor trusts the evaluation of "Mr. A”, “Mr. B” and “Mr. D” in the evaluation of the CD, but does not trust the evaluation of "Mr. C” at all. Performers are also very interested in trendy music. Based on this, how to use overlay processing Indicates what is being done.
- the filtering process by the filtering unit 4223 using a logical operation will be specifically described below. Excluded from the weighting process of spaces 2051, 2052, and 2054 that the results of the nodes included in the space 2053 that collected the untrusted 'Mr. C' recommendations and the space 2057 that collected the low-selling items were considered necessary. This is effective in cases such as In this case, the equivalent weighting process for the spaces 2051, 2052, and 2054 is performed, and An OR operation is performed with one or more existing as true, and those to which the result of the OR operation is applicable are excluded from the results of the equivalent weighting.
- the result of the OR operation of space 2053 and space 2057 is shown in FIG. Fig. 99 shows how to use this result to apply the filtering to the weighting score. In this case, those evaluated as true by the logical operation are excluded from the processing results.
- weight score 1 in Fig. 100, which is the addition of the weights of the spaces 2051, 2052, and 2054, is processed and analyzed from different viewpoints.
- the “rock” and “pops” in the “genre” property of the node are given priority, so the weight of the “rock” value is tripled and the “pops” value is used. Doubling! /
- FIG. 102 illustrates the case where projection by the projection unit 4411 is specifically used.
- projection is performed on the nodes 2360 and 2361 by the projection unit 4411 to form a multilingual node group.
- the nodes 2360 and 2361 have generated four nodes or node information.
- the projection transforms the space that can be expressed in two dimensions into three-dimensional information.
- the projected space is particularly called a "projected space”
- the projected node is called a "projected node”.
- the projection function is a process that treats multiple spaces as one space, and includes, for example, “single access of projective nodes”, “cooperation of node generation”, “cooperation of node deletion”, and “ It refers to “cooperation”, “access per projection space”, “access for each projection space of node”, “generation of space for each projection”, “deletion of space for each projection”, and “projection synchronization”. Projection logically has some or all of these functions.
- 2320, 2321, and 2322 are projection spaces, and 2301, 2302, and 2303 are the respective projected nodes.
- the "single access of a projection node" by the projection unit 4411 accesses a plurality of projected projection nodes as one in space. In other words, this means that the projected space as shown in Fig. 103B can be treated as accessing Fig. 103A.
- the "coordination of node generation" by the projection unit 4411 is to add a new node to another projection space when a new node is added. For example, when a node is added to the projective space 2320, a node is also created in the corresponding projective space 2321 or 2322.
- FIG. 103C illustrates cooperation of node generation by the projecting unit 4411. In this case, the projection nodes 2311, 2312, and 2313 are simultaneously generated.
- the "cooperation of node deletion" by the projection unit 4411 is to delete a node corresponding to another projection space when a node is deleted. For example, when the projection node 2301 is deleted, the projection nodes 2303 and 2305 are also deleted. Fig. 103B Fig. 103D shows the force that destroyed the projection nodes 2301, 2303, and 2305. In this case, three projective nodes are destroyed at the same time.
- the "access in units of projective space" by the projecting unit 4411 is such that each projective space can be treated as one independent space rather than a united space. That is, the projective spaces 2320, 2321, and 2322 can be treated as independent spaces.
- the "access for each projection space of a node" by the projection unit 4411 is to access a node by designating a projection space. For example, referring to FIG. 97B, if it is the projection space 2320, the projection node 2301 is specified, and if it is the projection space 2322, the node 2305 is specified.
- the access in units of projective space and the access in units of projective space of nodes are functions that enable the projective space to be treated as an independent space.
- the "deletion of space for each projection" by the projection unit 4411 is a function of destroying a space in units of a projection space. This means that as shown in FIG. 97G, the projection space 2320 is destroyed so that only the projection spaces 232 1 and 2322 are performed.
- This function is a function in which, when a space is projected, the projection is also applied to the synchronization of the nodes included in the space. For example, in the space of FIG. 104, the nodes 2371a and 2372a are synchronized. At this time, when projection is applied to 2371a, it means the function to apply projection to space 2372a. In FIG. 104, 2371b and 2372b are projected.
- projection nodes 2301, 2302, and 2303 are actually three nodes Is equivalent to treating a space as one.
- Link projection is a method of actually creating a plurality of link structures in a link space and controlling them by a program. This can be said to actually create multiple spaces so that they can be treated as projective spaces by the function of the program.
- Fig. 105B shows the result of link projection of Fig. 105A. As seen in Figure 105B, there are two spaces for each link configuration. In this case, a space composed of 2331a and 2332a and a space composed of 2331b and 2332b are generated as elements of projection.
- the node projection is a process of creating a uniform structure for each node constituting a space, and is a method of creating a plurality of each node and controlling them by a program.
- Node projection of Fig. 105A. Fig. 105C. This is a structure in which a plurality of nodes or node information is added to each node. In this case, node 2331 is converted to node 2331c, and node 2332 is converted to node 2332c. That is, each node holds the projected node information.
- Node projection is easier to implement than other methods for projecting in a space that is not composed of links, such as an intangible space, and for synchronous projection.
- FIG. 106 shows a case where the node projection is created by a link list.
- Node 2380 is a list of projected nodes or node information 2381, 2382, and 2383.
- This link can be a list structure in memory or an instance of an object, or it can be a tag language such as XML or HTML. This link may be the technology used to link nodes.
- Fig. 107 illustrates an implementation for linking a node to a data group in node projection.
- the data group of node 2391 is connected to node 2390. This is in memory If implemented, it is better to store the projected node or node information in the data container. When saving data to a disk storage device or the like, it is better to save the data in a file format or in a binary-compressed form. The point is, even if the inside is random, it is only necessary to be able to access it as a projected node!
- the data addition projection is a method of connecting projection data to a space.
- Fig. 105 5A is the result of data-added projection.
- Fig. 105D This includes those that use layers to achieve projection.
- the projection is realized by using the additional information 2333 for the projected space.
- the interior of 2333 contains information about the space. Using this information, it is treated as if there are multiple spaces.
- the data addition projection will be described with reference to FIG. Figure 108 projects the space consisting of nodes 2401, 2402, and 2403.
- Node 2401 has 2404 forces.
- Node 2402 has 2405 forces.
- Node 2403 has 2406 forces.
- the information for projection may be stored in 2404, 2405, and 2406.
- the design may be such that the common resources 2407 are referred to as a whole 2404, 2405, and 2406. In this way, projection is performed by adding data to the space.
- the independent space projection by the projection unit 4411 is a method of creating a type of independent space node and holding a plurality of pieces of spatial information inside or outside to implement projection.
- Figure 105E shows the independent space projection of Figure 105 ⁇ . As long as the data for projection inside the independent space node 2334 can be handled by the projection function, a plurality of spaces may be held inside in any form, or may be expressed as a completely different format.
- the independent space projection will be described with reference to FIG.
- the space definition 2418 is connected to the independent space node 2411 externally or internally.
- 2418 contains three projective spaces, 2412 and 2413 space, 2414 and 2415 space, 2416 and 2417 space.
- Figure 110A is a space that represents a document.
- the space 2481 is related to important documents and needs to be created in multiple languages. In this case, do not use projection. Otherwise, the space design is as shown in Fig. 110B.
- Figure 110B many spaces are created for each language version. For example, 2481 is for Japanese, 2482 is for English, and 2483 is for French. Thus, a new space is created for each different version. This means that the space increases with each version.
- the use of projection does not logically increase the space itself as shown in Fig. 110C.
- each of 2481a, 2481b, and 2481c is a space for one language. Creating a different space for each different version, as shown in Figure 110B, will increase management costs.
- Space is a composition that expresses or collects a number of concepts. When changing the concept to be expressed, it is necessary to change the configuration of the space. At this time, if there are simply multiple spaces for each version, it will be necessary to apply changes and additions to all spaces. This is a problem that arises because the concepts to be expressed are the same, but they are implemented as different spaces with only the difference in the format, and these are very expensive to maintain.
- the concept to be expressed is dispersed as a different space only by the difference of the expression form, which makes it difficult for a third party to understand the relationship between the concept and the space.
- the projection that connects the concepts to be expressed into one is an important function.
- FIG. 111 is a diagram for describing the projection unit 4411 in detail.
- a functional block diagram of a general program that realizes projection is shown. These functions are implemented by one or more programs. Further, the program may be processed by a plurality of computers.
- the space corresponding to the projection and the processing are linked.
- the spatial data section 4201 is controlled by the projection section 4411.
- the spatial data section 4201 or the projecting section 4411 can be appropriately accessed through the interface to the space 4413, so even if the inside of the spatial data section 4201 is in a special format different from the normal space, it is easy. And secure access.
- the data of the projected space exists as a link projection 4430, a node projection 4431, a data addition projection 4432, or an independent space projection 4433. Although these data may be accessed directly, it is better to operate through the interface to space 4413 because the risk of data destruction without having to consider complex internal configurations and specifications is reduced.
- the projecting unit 4411 includes “single access of projective nodes”, “cooperation of node generation”, “cooperation of node deletion”, “cooperation of space transformation”, “access in projective space units”, “node It performs the function of projecting by performing operations such as “access for each projection space”, “generation of space for each projection”, “deletion of space for each projection”, and “projection synchronization”.
- Layers can be applied to nodes, links, and spaces.
- a layer is a structure in which information is defined in parallel with a node or space, and has a structure in which additional information is given in parallel with a node that does not attach information to the node itself.
- To be defined in parallel means to be defined in a form that corresponds to nodes and spaces.
- Various information can be added to the space by using layers.
- nodes do not operate themselves, making it easy to add information and reducing the risk of data corruption.
- it can be used in various situations by stacking multiple layers or combining them depending on the situation. Layers can separate things that are different from the concept that the node or space wants to represent.
- all information that is different from the concept you want to express should be included in the layer.
- Layers can be applied to a single node or space, are flexible, and are very useful when adding information to a space. It is especially effective to use security, events, and priorities.
- the layer can give information to each node in the space to which the layer is added, or can give information as the whole space.
- the layers include a "structural layer” and a "substance layer".
- the structural layer adds a layer to the spatial structure itself, and applies a layer to the node structure.
- the entity layer adds a layer to the entity itself, that is, performs the operation on the node itself. In other words, the effect of the layer is applied to all nodes that have the same synchronization point for the node to which the substantive layer is added.
- the structure layer and the substance layer will be described with reference to FIG. Nodes 3141 and 3142 are synchronized with each other. At this time, even if a structural layer is applied to a node in the space 3140, a layer is added to the node 3141. No layer is added to the node 3142. However, if the physical layer is added to the nodes in the space 3140, the physical layer is applied to the node 3142 as well as the node 3141.
- the structure layer will be described with reference to FIG. In Figure 112, one space and two structural layer forces S are defined. Nodes 3100, 3101, 3102, 3103, and 3104 form a link space.
- a structure layer is composed of 3110, 3111, 3112, 3113 and 3114, 3115, 3116, 3117.
- two structural layers are connected to the space.
- the structural layer it is possible to define a layer for the link, but in this case, the structural layer is added only to the nodes in the space.
- the structure layers 1 and 3114, 3115, 3116 and 3117 are respectively formed in layers 3110, 3111, 3112 and 3113. The straight broken line shown in FIG.
- 112 indicates the correspondence between the node and the structure layer, and the circular broken line indicates the structure layer.
- additional information is given to the structure layer, and various processes are performed inside or outside the program based on the additional information. Layers do not need to be composed of links like nodes, even when layers are added to a force link space that exists in parallel with nodes.
- the entity layer will be described with reference to FIG. Node 3150 and node 3152 are synchronized with each other. Nodes 3150, 3151, and 3152 have physical layer power. At this time, node 3153 and node 3150 and node 3152 force in layer 1 and node 3151 and 3154 force S are connected. Similarly, node 3150 and 3152 are connected to layer 3155, and node 3151 is connected to 3156. In this way, the entity layer is tied to the node entity, not the spatial structure.
- Layer is a force that can exist outside or inside a node.Layer logically gives information to an external force node, so it is desirable to hold a lot of information outside even in implementation. .
- the layer can be implemented on a computer-readable recording medium or on a network. That is, it can be implemented in any environment where the node can be realized. It is in memory It can be an entity, or it can be XML-based on disk.
- FIG. 114 shows a method described in the node itself.
- Reference numeral 3150 denotes a node or reference information to the node. 3150 stores information on layers. In addition, information for the node is stored in layer 1 3151.
- FIG. 115 shows a case where the relationship between nodes and layers is described externally. 3155 is a node
- information is added to a node as asynchronous information addition, and the physical layer is usually treated as a synchronous information addition to a node.
- the part is the definition of the structural layer. In a real common implementation, this tag would contain more information or a reference to the information.
- the layer self-expanding function refers to a function in which a layer is applied to a node to which a layer is applied, or to an added portion when a space is added. Furthermore, it is also effective to use a function of selecting a space link and a function of spatial filtering for this function to implement a self-expanding space in a more flexible manner. This function is mainly applied in the structural layer.
- the self-expanding function of the layer will be described with reference to the drawings.
- the nodes 3120, 3121, and 3122 in FIG. 120A form an empty space at the nodes! 3130, 3131 and 3132 layers are formed.
- follow-up node 3123 to 3122 As shown in FIG. 120B, follow-up node 3123 to 3122. Then, layer 3133 of FIG. 12 OC is expanded.
- Introducing a security layer in implementing security can make security management more manageable and easier. This can be achieved by mapping security layers into space. Normally, the security layer contains one or more user information and security right information. As a result, security for each user and the correspondence between the authority and space of multiple users become clearer and easier to manage. Note that the security layer is a layer in which security information is added.
- the security layer will be described with reference to FIG. 3215 is a security layer for user 3201, and 3216 is a security layer for user 3202. Each security layer masks part of the space. Instruction 3202 for user 3201 is excluded by security layer 3215. Instruction 3222 is applied to 3211 without being affected by the security layer. Similarly, instruction 3221 applies to node 3210 and instruction 3223 has been excluded.
- Fig. 122 shows an example of realizing security by the security layer using a step diagram.
- step S401 an operation on the space is detected.
- step S402 it is checked whether a security layer exists in the space to be operated.
- the security set in step S403 is analyzed to determine whether to permit the operation.
- S404 is executed when the operation is permitted, S405 is not permitted, the operation is invalidated, and the notification process is performed at the same time. Be executed.
- an event layer in which event information is layered is introduced, so that the event can be handled more flexibly.
- Events basically consist of “reaction” and “notification”.
- the event layer should record information such as the ability to react to what and to whom and how to communicate, and implement the function of the program event.
- An event layer is a layer to which event information for notifying internal or external users of changes in space, commands to space, or temporal changes is added.
- FIG. 123 shows an example of realizing an event using an event layer by using a step diagram.
- Step 501 means that an event such as an operation or change to the space occurs.
- step 502 it is checked whether or not the event layer corresponding to the space to be operated or the changed space exists.
- step S503 the force of the event information corresponding to the operation or change is confirmed.
- step S504 a notification process is performed based on the event information.
- steps S502 and S503 information about the event is collected, and in S504, an actual notification is performed.
- the space in the figure represents a software bug report for a company.
- a security layer is set in this space.
- security information is added using the structural layer 3251 to prohibit changes by outsiders in the entire space.
- security information should be set in 3252 as a structural layer that makes the creator-specific part invisible to outsiders.
- FIG. 125 shows a functional block diagram of a general program that realizes a security layer. These functions are implemented by one or a plurality of programs. Further, the program may be processed by a plurality of computers.
- the user 4802 accesses the spatial data section 4201 through the security gateway 4803 to! /.
- the security gateway 4803 accesses the security layer information 4804 and permits operation according to the security level.
- the automatic classification processing unit 4809 is a processing unit that expands a space by adding a node that configures the space, and obtains a node or information associated with the node, and starts an automatic classification processing start unit. 4810 and an automatic classification unit 4811.
- the automatic classification unit 4811 includes a traversing unit 4812 that links the links of the nodes.
- the determining unit 4813 that determines whether the power is appropriate and the determining unit 4813 determine that When it is determined that the node is appropriate, it is composed of an adding unit 4814 for adding a node, and the determining unit 4813 is further composed of an existence / non-existence condition determining unit 4820 and a traversing possibility / non-permission condition determining unit 4821. Further, it comprises an individual adding section 4830, a synchronization adding section 4831, and a link adding section 4832. The outline of these functions will be described with reference to FIG. 134 of a functional block diagram of a general program that implements automatic classification processing. These functions are implemented by one or more programs. Further, the program may be processed by a plurality of computers.
- the "automatic classification processing" of the space by the automatic classification processing unit 4809 facilitates extension and addition to the link space.
- This function can add new nodes by using existence conditions that are connected to nodes inside or outside the nodes while traversing the space. Judge whether it is a good one.
- the general implementation of the existence condition expression is generally considered to be able to be described in a script language or programming language like a node judgment expression using information of newly added nodes. It is.
- One existence condition expression may correspond to a plurality of nodes.
- FIG. 126A shows a node to which a presence / absence conditional expression is added.
- Nodes 3610, 3611, 3612, and 3613 have presence / absence conditional expressions 3620, 3621, 3622, and 3623, respectively.
- FIG. 126B shows the result of the newly added node.
- the newly added nodes were determined to be inappropriate according to the conditional expressions 3620 and 3621.
- the new nodes 3630 and 3631 were permitted to be added as appropriate. .
- FIG. 127 is shown as an example of a rough flow of the automatic classification process, and will be described step by step.
- the traversing unit 4812 checks whether or not there is a node to be deleted, and terminates the process if there is no such node.
- the next node is entered.
- the existence / non-existence condition determination unit 4820 determines whether or not the existence is possible using the existence / non-existence condition expression. If it is determined that the existence is possible, the addition unit 4814 adds the node to the node in S604. Further, the flow of FIG. 127 is recursively executed inside S602 or S604.
- FIG. 12 illustrates a case where a node is added at the same time as the presence / absence conditional expression in the automatic classification process. This can be realized by replacing S604 in the flowchart of FIG.
- the traveling permission / non-permission condition determination unit 4821 controls a node to be searched in the automatic classification process using a traveling permission / prohibition condition expression.
- the automatic classification process starts from the root node 3640.
- the patrol processing 3650 reaches the node 3642, if the patrol permission / non-permission condition determination unit 4821 determines that the patrol is not possible by the patrol permission / prohibition conditional expression set in the conditional expression 3652, the packet does not reach the nodes 3643 and 3644. The process ends.
- synchronous automatic addition by the synchronous addition unit 4831 means that nodes added after the second are added in synchronization with the node added first.
- the existence / non-existence conditional expression and the traversability / non-existence conditional expression may be linked to the node by a mechanism using either "synchronous information addition” or "spatial information addition”. By using synchronous information addition, conditional expressions can be shared between synchronized nodes.
- the existence / non-existence conditional expression and the tour availability / non-existence conditional expression need not correspond to one node. That is, these expressions may be shared by a plurality of nodes.
- the major flow of the processing by the automatic classification processing unit 4809 is that the data classification is received from the automatic classification processing start unit 4810, the automatic classification unit 4811 is executed, and the spatial data The node is added to the data section 4201.
- Automatic classification processing start unit 4810 acquires or generates data for starting automatic classification processing. This part may be based on an external or internal program, when a change outside the program is sensed, or input by a human.
- the automatic classification unit 4811 includes a traveling unit 4812, a determination unit 4813, and an addition unit 4814.
- the traversing unit 4 812 traverses the nodes of the spatial data unit 4201, and when the determining unit 4813 determines that the node can be added, the adding unit 4814 adds the node to the spatial data unit 4201.
- the traveling unit 4812 acquires the information of the spatial data unit 4201 and performs a traveling around the space or a process having the same effect as the traveling. While circulating in the circulating unit 4812, the decision unit 4813 determines whether to add a node or whether to traverse the node further.
- the determination unit 4813 determines the presence / absence condition determination unit 4820 that determines whether to add the presence permission conditional expression associated with the node and determines whether to add the node, and determines whether or not to perform the tour by determining the traversal permission / non-permission condition expression. It comprises a patrol permission / non-permission condition determination unit 4821.
- the existence permission condition expression and the traversability condition expression used in the judgment unit 4813 are usually associated with a node using a layer or a similar mechanism as shown in 4815.
- the existence / non-existence conditional expression 'circability / non-existence conditional expression 4815' may be a script 'program language 4841 which is usually defined by an external program 4842 or the like. Actually, if the conditional expression is complicated, it seems to be common to call the function of the internal or external program from Script's programming language 4841 and process it.
- the adding unit 4814 performs a process of adding a new node to the spatial data unit 4201. Individual type automatic adding unit is realized by using individual adding unit 4814, synchronous automatic adding unit is used by using synchronous adding unit 4831, and link type automatic adding unit is realized by using link adding unit 4832. I do.
- spatial storage With the implementation of the space in the spatial data section 4201 as an independent storage, because it facilitates data handling, ease of processing, and coordination of processing.
- semantic storage it is possible to link with a relational database or XML database to improve performance, centralize data management, and coexist with existing information systems. Become. This is based on the technology of distributed management of space.
- the definition of spatial storage is storage that can use spatial overlay, projection, layer-based, or automatic classification.
- 3802 is a spatial storage
- 3801 is a client of the spatial storage
- 3803 is a relational database or XML database.
- the communication between 3801, 3802, and 3803 is performed by 3804 and 3805.
- the communication of 3804 and 3805 may be the communication method used in the distributed management of space.
- the most powerful function in using the space is a process using the superposition of the space by the superposition calculation processing unit 4211. This makes it possible to easily convert ambiguous data into more semantically detailed information.
- the most effective use would be in technical databases such as state-of-the-art technology, medicine, law, and information science, as well as for searching and exchanging knowledge on the Internet. Compared to other information management methods, the more the vague and complicated data is handled, the greater the effect of the present invention. Also, since human emotional information can be used, it is also effective for art databases such as paintings, movies, and music information.
- a supervising search can be performed by superimposing a space by the superimposition calculation processing unit 4211, a learning search device of an educational institution may be used.
- the present invention is applied to the Internet directory search engine.
- the mainstream Internet search engines are robot-type search engines used by google and others.
- relatively selective sites are registered in the directory search engine, and it can be said to be an effective search engine when searching for a specific set of information.
- Typical directory-type search engines include Yahoo categories and google directories.
- FIG. 152 shows a search engine system to which the present invention is applied.
- 7111, 7112, and 7113 represent users who use the search engine.
- the user connects to the search engine device 7101 through the information communication network 15 using a Web browser such as Internet Explorer or another communication program.
- the information provision interface unit 7102 implements the interface function as a search engine, and stores the address and name of the Web site in the spatial data unit. If it is necessary to perform the superimposition process, the calculation is performed using the information space processing device 4000, and the result is delivered to the user through 7102.
- a directory-type search engine categorizes Web sites into categories by people or computers to create tree-like information, and searchers can find information by following the categories.
- FIG. 153 shows an example of the categorization of the directory type search engine partially omitted.
- a category that is usually useless Rise is also useful classification information.
- the space is classified according to the registration date of the Web site or the corresponding language.
- the ivy space alone is almost worthless as the classification information of the Web page, but it becomes meaningful information if the overlay processing is used.
- ivy information is not very valuable on its own, so it is not created by website classification.
- such information is effective value-added information for a system that uses overlay processing.
- the use of the search engine system 7100 also makes it easy to search using constantly changing information. For example, consider the case of using Web page access ranking. Fig. 154 shows the access ranking information in space 7131. Such information changes over time. In addition, since the data is of interest and not very meaningful on its own, it is difficult to categorize it with ordinary index-type search engines. However, if an access ranking space is created, a search that adds value, such as “Web site of the power category you are looking for but has an access ranking of 1 to 2000”, can be easily performed by overlay processing. As a result, it is possible to narrow down Web sites that are likely to be the source of information requested by users.
- the expansion of space does not basically affect the nodes. Therefore, the space can be expanded relatively easily in the search engine system 7100 as well. Therefore, by creating and adding spaces from various perspectives, it is possible to significantly increase the number of web page search results. Also, even if a space is created with subjective information, if the user can understand it, it is possible to use information search by superposition.
- FIG. 155 is a drawing in which a new space 7132 is added to the space of FIG. 154 created in the space data section 4201.
- This space 7132 lists the famous sites of famous people called Opinion Leaders.
- the genre you are looking for, but the access ranking is 1st to 2000th by adding a superimposition with the space 7133, A site with an access ranking of 1-2000 and recommended by celebrity A. " A more specialized search like this can be performed.
- a more specialized search like this can be performed.
- the user can search for information on the Internet from various perspectives by using the superposition process while reducing or reducing various spaces.
- a method of using the present invention for a bulletin board system will be described below using an actual example.
- a bulletin board system is a system in which many people share information by writing information to the bulletin board through an information communication network such as the Internet.
- the bulletin board system 7300 in FIG. 156 is a bulletin board system in which a search system is enhanced by using the present invention.
- 7311, 7312, and 7313 represent bulletin board users who use the bulletin board.
- Bulletin board system The 7300 links and manages bulletin board information and spatial information.
- the user connects to the bulletin board device 7301 through the information communication network 15 using a Web browser such as the Internet Explorer or another communication program.
- the user transfers arbitrary information to the bulletin board device 7301 for discussion and information exchange.
- bulletin board thread In a general bulletin board system, one bulletin board is created for each topic. This is usually called a bulletin board thread.
- the thread is composed of multiple message boards that write messages.
- Figure 157 is a simple representation of a bulletin board.
- 7321 means a thread which is a unit of topic of a bulletin board
- 7322 means a message board.
- FIG. 158 shows a tree-type bulletin board. In this case, a response can be returned to each arbitrary message board.
- message board 7327 is a response message board for 7326.
- a message boat is used as a node and a bulletin board.
- the signboard in FIG. 157 can be represented as in FIG.
- the message board 732 corresponds to the node 7323
- the message board 7323 corresponds to the node 7333.
- the tree-type bulletin board in FIG. 158 can be represented as in FIG.
- the message board 7326 corresponds to the node 7336
- the message board 7328 corresponds to the node 7338!
- the weighting cost by comparison and evaluation that makes it easier to handle text information etc. is much higher than that by equivalence evaluation. Therefore, when implementing as a system, the text information on the message board is replaced with information for search processing such as vector information, and the information is stored in the spatial data section 4201 and compared with that information. The method of performing weighting and the like is common.
- the thread may become too large due to the incandescence of discussions on the bulletin board or the subdivision of the contents. In such a case, if the comparative examination weighting process is performed by the thread alone, an important message board can be emerged as a whole.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Description
Claims
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2006514606A JPWO2005122014A1 (ja) | 2004-06-14 | 2005-06-14 | 情報空間処理装置、プログラム、および、方法 |
JP2007523391A JPWO2007000893A1 (ja) | 2005-06-14 | 2006-06-14 | 情報空間処理装置、プログラム、および、方法 |
PCT/JP2006/311903 WO2007000893A1 (ja) | 2005-06-14 | 2006-06-14 | 情報空間処理装置、プログラム、および、方法 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2004-175714 | 2004-06-14 | ||
JP2004175714 | 2004-06-14 |
Publications (2)
Publication Number | Publication Date |
---|---|
WO2005122014A2 true WO2005122014A2 (ja) | 2005-12-22 |
WO2005122014A1 WO2005122014A1 (ja) | 2005-12-22 |
Family
ID=
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2007000893A1 (ja) * | 2005-06-14 | 2007-01-04 | Mitsuhiro Tanada | 情報空間処理装置、プログラム、および、方法 |
JP2008225997A (ja) * | 2007-03-14 | 2008-09-25 | Nippon Telegr & Teleph Corp <Ntt> | メタデータ管理方法、メタデータ管理システム、及び、メタデータ管理プログラム |
JPWO2007000893A1 (ja) * | 2005-06-14 | 2009-01-22 | 充弘 店田 | 情報空間処理装置、プログラム、および、方法 |
US8818930B2 (en) | 2009-05-18 | 2014-08-26 | Takatoshi Yanase | Knowledge base system, logic operation method, program, and recording medium |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2007000893A1 (ja) * | 2005-06-14 | 2007-01-04 | Mitsuhiro Tanada | 情報空間処理装置、プログラム、および、方法 |
JPWO2007000893A1 (ja) * | 2005-06-14 | 2009-01-22 | 充弘 店田 | 情報空間処理装置、プログラム、および、方法 |
JP2008225997A (ja) * | 2007-03-14 | 2008-09-25 | Nippon Telegr & Teleph Corp <Ntt> | メタデータ管理方法、メタデータ管理システム、及び、メタデータ管理プログラム |
US8818930B2 (en) | 2009-05-18 | 2014-08-26 | Takatoshi Yanase | Knowledge base system, logic operation method, program, and recording medium |
Also Published As
Publication number | Publication date |
---|---|
JPWO2005122014A1 (ja) | 2008-04-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Baca | Introduction to metadata | |
Aggarwal et al. | Text mining in social networks | |
US8290923B2 (en) | Performing large scale structured search allowing partial schema changes without system downtime | |
US20080104032A1 (en) | Method and System for Organizing Items | |
Salatino et al. | Improving editorial workflow and metadata quality at springer nature | |
Khan et al. | Weblens: Towards interactive large-scale structured data profiling | |
Xu et al. | Application of rough concept lattice model in construction of ontology and semantic annotation in semantic web of things | |
Wei et al. | DF-Miner: Domain-specific facet mining by leveraging the hyperlink structure of Wikipedia | |
Collarana et al. | Fuhsen: A federated hybrid search engine for building a knowledge graph on-demand (short paper) | |
Za et al. | Knowledge creation processes in information systems and management: Lessons from simulation studies | |
Jupp et al. | A flexible API and editor for SKOS | |
Priss | Facet-like structures in computer science | |
Ke et al. | Structure and pattern of social tags for keyword selection behaviors | |
Nešić | Semantic document model to enhance data and knowledge interoperability | |
Zhang et al. | An exploratory study on utilising the web of linked data for product data mining | |
WO2007000893A1 (ja) | 情報空間処理装置、プログラム、および、方法 | |
Signore | The semantic web and cultural heritage: Ontologies and technologies help in accessing museum information | |
Stoilos et al. | An ontology-based interactive system for understanding user queries | |
Kurki et al. | Authority control of people and organizations on the semantic web | |
Wei et al. | ExNa: an efficient search pattern for semantic search engines | |
Mimouni et al. | Answering complex queries on legal networks: A direct and a structured IR approaches | |
WO2005122014A2 (ja) | 情報空間処理装置、プログラム、および、方法 | |
Stegemann et al. | Interactive construction of semantic widgets for visualizing semantic web data | |
CN101452462A (zh) | 用于自动建立网络上信息对象之间层次结构的方法和系统 | |
Chen et al. | A bag-of-semantics model for image clustering |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A2 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BW BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE EG ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KM KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NA NG NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SM SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW |
|
AL | Designated countries for regional patents |
Kind code of ref document: A2 Designated state(s): GM KE LS MW MZ NA SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LT LU MC NL PL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
NENP | Non-entry into the national phase in: |
Ref country code: DE |
|
WWW | Wipo information: withdrawn in national office |
Country of ref document: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2006514606 Country of ref document: JP |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 05751172 Country of ref document: EP Kind code of ref document: A2 |