Terminal, data processing device, terminal control program, data processing program, terminal control method, and data processing method

Abstract

To prevent data possessed by an organization or the like from being compared with data obtained by subjecting abstraction processing to the aforementioned data.SOLUTION: A terminal generates abstraction partial data, first abstraction anchor data and second abstraction anchor data by an abstraction function, discards the abstraction function, and transmits these three data and label data to a data processing device. The data processing device integrates a plurality of abstraction partial data to generate integrated data, integrates a plurality of second abstraction anchor data to generate integrated second anchor data, generates an integrated analysis model for outputting label data showing a label corresponding to contents shown by an element included in the integrated data on the basis of the integrated data, and generates second label data by using the integrated second anchor data and the integrated analysis model to transmit the second label data to the terminal. The terminal uses the second anchor data and the second label data, and generates an integrated analysis model showing a relation between contents shown by each of the two data.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a terminal, a data processing device, a terminal control program, a data processing program, a terminal control method, and a data processing method. [Background technology]

[0002] In recent years, the networking of society has progressed rapidly, creating an environment for collecting data in various fields of society. For example, various institutions, such as medical institutions and financial institutions, independently collect and store large amounts of data. Against this background, there are hopes for improving productivity through technology that integrates this data and analyzes it as big data while protecting personal information, trade secrets, etc. One example of such technology is a distributed data integration device disclosed in Patent Document 1.

[0003] This distributed data integrating device includes an acquiring unit, an anchor data conversion unit, a calculating unit, and an analysis target data conversion unit. The acquiring unit acquires, for each piece of analysis target data, an anchor data intermediate representation which is an intermediate representation obtained by converting anchor data, which is data commonly used in integrating multiple pieces of distributed analysis target data, using a first function, and an analysis target intermediate representation which is an intermediate representation obtained by converting the analysis target data using the first function. The anchor data conversion unit converts the multiple anchor data intermediate representations acquired by the acquiring unit for each piece of analysis target data using a second function. The calculation unit calculates a second function for each piece of analysis target data that minimizes the difference between the anchor data intermediate representations converted by the anchor data conversion. The analysis target data conversion unit converts the analysis target intermediate representation acquired by the acquiring unit for each piece of analysis target data using the second function calculated by the calculation unit. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] International Publication No. 2020 / 137728

[0005] However, since the above-mentioned distributed data integration device is capable of comparing the original data with the intermediate representation to be analyzed, it cannot completely eliminate the possibility that the content indicated by the original data may become known to persons other than the institution that holds the original data. Summary of the Invention [Problem to be solved by the invention]

[0006] The present invention has been made in consideration of the above circumstances, and aims to provide a terminal, a data processing device, a terminal control program, a data processing program, a terminal control method, and a data processing method that make it impossible to compare data held by an institution or the like with data that has undergone abstraction processing. [Means for solving the problem]

[0007] One aspect of the present invention includes an abstraction function generation unit that generates an abstraction function used to perform abstraction processing on partial data, first anchor data, and second anchor data; an abstraction execution unit that applies the abstraction function to the partial data to generate abstracted partial data, applies the abstraction function to the first anchor data to generate first abstracted anchor data, and applies the abstraction function to the second anchor data to generate second abstracted anchor data; an abstraction function discarding unit that discards the abstraction function after the abstracted partial data, the first abstracted anchor data, and the second abstracted anchor data have been generated; The terminal includes a data sending unit that sends one abstraction anchor data, the second abstraction anchor data, and label data to a data processing device; a data receiving unit that receives from the data processing device second label data that indicates labels corresponding to content indicated by elements included in integrated second anchor data generated by integrating multiple second abstraction anchor data using multiple first abstraction anchor data; and a model generating unit that uses the second anchor data and the second label data to generate an integrated analysis model that indicates the relationship between the content indicated by the second anchor data and the content indicated by the second label data.

[0008] In addition, in the above-mentioned terminal, the abstraction function discarding unit discards the abstraction function before the abstraction partial data, the first abstraction anchor data, and the second abstraction anchor data are transmitted to the data processing device.

[0009] In the above-described terminal, the abstraction function generation unit generates the abstraction function by at least one of a process of randomly extracting a part of the partial data and a process of adding a random element to the abstraction function.

[0010] In the above-described terminal, the abstraction execution unit randomly interchanges the rows of the abstract partial data and the label data equally.

[0011] One aspect of the present invention is a data processing device comprising: a data receiving unit that receives abstraction partial data, first abstraction anchor data, second abstraction anchor data, and label data for each terminal; a label data generating unit that generates integrated data by integrating multiple pieces of abstraction partial data using the first abstraction anchor data received for each terminal, generates integrated second anchor data by integrating multiple pieces of second abstraction anchor data using the first abstraction anchor data received for each terminal, generates an integrated analysis model that outputs label data indicating labels corresponding to content indicated by elements included in the integrated data based on the integrated data, and generates second label data indicating labels corresponding to content indicated by elements included in the integrated second anchor data using the integrated second anchor data and the integrated analysis model; and a data transmitting unit that transmits the second label data to the terminal.

[0012] One aspect of the present invention includes an abstraction function generation function that generates an abstraction function used to perform abstraction processing on partial data, first anchor data, and second anchor data; an abstraction execution function that applies the abstraction function to the partial data to generate abstracted partial data, applies the abstraction function to the first anchor data to generate first abstracted anchor data, and applies the abstraction function to the second anchor data to generate second abstracted anchor data; an abstraction function discarding function that discards the abstraction function after the abstracted partial data, the first abstracted anchor data, and the second abstracted anchor data have been generated; and an abstraction function discarding function that discards the abstraction function after the abstracted partial data, the first abstracted anchor data, and the second abstracted anchor data have been generated. The terminal control program causes a terminal to execute a data transmission function that transmits car data, the second abstraction anchor data, and label data to a data processing device; a data reception function that receives from the data processing device second label data that indicates labels corresponding to content indicated by elements included in integrated second anchor data generated by integrating multiple second abstraction anchor data using multiple first abstraction anchor data; and a model generation function that uses the second anchor data and the second label data to generate an integrated analysis model that indicates the relationship between the content indicated by the second anchor data and the content indicated by the second label data.

[0013] One aspect of the present invention is a data processing program that causes a data processing device to execute: a data receiving function that receives abstraction partial data, first abstraction anchor data, second abstraction anchor data, and label data for each terminal; a label data generation function that generates integrated data by integrating multiple pieces of abstraction partial data using the first abstraction anchor data received for each terminal, generates integrated second anchor data by integrating multiple pieces of second abstraction anchor data using the first abstraction anchor data received for each terminal, generates an integrated analysis model that outputs label data indicating labels corresponding to content indicated by elements included in the integrated data based on the integrated data, and generates second label data indicating labels corresponding to content indicated by elements included in the integrated second anchor data using the integrated second anchor data and the integrated analysis model; and a data transmission function that transmits the second label data to the terminal.

[0014] One aspect of the present invention is A terminal control method executed by a computer, the computer comprising: An abstraction function generating unit or an abstraction function generating function generates an abstraction function to be used for performing abstraction processing on the partial data, the first anchor data, and the second anchor data; The computer an abstraction execution unit or an abstraction execution function applies the abstraction function to the partial data to generate abstract partial data, applies the abstraction function to the first anchor data to generate first abstract anchor data, and applies the abstraction function to the second anchor data to generate second abstract anchor data; The computer an abstraction function discarding unit or an abstraction function discarding function discards the abstraction function after the abstraction partial data, the first abstraction anchor data, and the second abstraction anchor data are generated; The computer a data transmitting unit or a data transmitting function transmitting the abstraction partial data, the first abstraction anchor data, the second abstraction anchor data, and the label data to a data processing device; The computera data receiving unit or a data receiver receives, from the data processing device, second label data indicating labels corresponding to contents indicated by elements included in integrated second anchor data generated by integrating a plurality of the second abstract anchor data using a plurality of the first abstract anchor data; The computer A terminal control method in which a model generation unit or a model generation function uses the second anchor data and the second label data to generate an integrated analysis model that shows the relationship between the content indicated by the second anchor data and the content indicated by the second label data.

[0015] One aspect of the present invention is A data processing method executed by a computer, the computer comprising: a data receiving unit or a data receiving function receiving abstraction partial data, first abstraction anchor data, second abstraction anchor data, and label data for each terminal; The computer a label data generation unit or a label data generation function generates integrated data by integrating a plurality of the abstraction partial data using the first abstraction anchor data received for each of the terminals, generates integrated second anchor data by integrating a plurality of the second abstraction anchor data using the first abstraction anchor data received for each of the terminals, generates an integrated analysis model that outputs label data indicating labels corresponding to content indicated by elements included in the integrated data based on the integrated data, and generates second label data indicating labels corresponding to content indicated by elements included in the integrated second anchor data using the integrated second anchor data and the integrated analysis model; The computer The data processing method includes transmitting the second label data to the terminal by a data transmitting unit or a data transmitting function. [Effects of the Invention]

[0016] According to the present invention, it is possible to provide a terminal, a data processing device, a terminal control program, a data processing program, a terminal control method, and a data processing method that make it impossible to compare data held by an institution or the like with data that has undergone abstraction processing. [Brief explanation of the drawings]

[0017] [Figure 1] FIG. 1 is a diagram illustrating an example of an integrated analysis model generation system according to an embodiment of the present invention. [Figure 2] FIG. 10 is a diagram showing an example of overall data according to an embodiment of the present invention. [Figure 3] FIG. 10 is a diagram illustrating an example of partial data according to the embodiment of the present invention. [Figure 4] FIG. 2 is a diagram illustrating an example of a functional configuration of a terminal according to an embodiment of the present invention. [Figure 5] FIG. 1 is a diagram illustrating an example of a functional configuration of a data processing device according to an embodiment of the present invention. [Figure 6] FIG. 4 is a sequence diagram showing an example of processing executed by a terminal and a data processing device according to an embodiment of the present invention. [Figure 7] FIG. 10 is a diagram showing an example of the relationship between the number of institutions and the accuracy rate when abstracted partial data indicating corporate ratings is integrated and analyzed using an integrated analysis model generated by a terminal according to an embodiment of the present invention. [Figure 8] FIG. 10 is a diagram showing an example of the relationship between the number of positions and the accuracy rate when abstracted partial data representing handwritten numbers is integrated and analyzed using an integrated analysis model generated by a terminal according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0018] [Embodiment] First, an overview of an integrated analytical model generation system according to an embodiment will be described with reference to Fig. 1 to Fig. 3. Fig. 1 is a diagram showing an example of an integrated analytical model generation system according to an embodiment of the present invention. As shown in Fig. 1, the integrated analytical model generation system 1 includes a terminal 10-1, ..., a terminal 10-c (c: a natural number equal to or greater than 2), and a data processing device 20.

[0019] The terminals 10-1, ..., 10-c and the data processing device 20 are all connected to the network NW shown in Fig. 1 and are capable of communicating with each other. The network NW is, for example, the Internet, a WAN (Wide Area Network), an intranet, or a LAN (Local Area Network).

[0020] Terminals 10-1, ... and terminal 10-c are computers, such as servers, managed by different organizations, institutions, etc. The institutions referred to here are, for example, corporate rating agencies or medical institutions. Data processing device 20 is a computer, such as a server, that collects and analyzes partial data from terminals 10-1, ... and terminal 10-c.

[0021] The terminals 10-1, . . . and 10-c receive partial data X1, . . . and partial data X2, . . . which are parts of the entire data, respectively. c Partial data X1, ... and partial data X c Each of these indicates information that must be kept secret by the organization that manages the terminal 10-1, . . . or the terminal 10-c.

[0022] Fig. 2 is a diagram showing an example of overall data according to an embodiment of the present invention. For example, as shown in Fig. 2, the overall data has a structure in which multiple feature amounts are defined for each sample, and is expressed as a matrix. For example, if the overall data is data related to the financial aspects of a company, these feature amounts are indices representing features related to the company's rating. Alternatively, if the overall data is data related to handwritten numeric imagery depicting handwritten numeric characters, these feature amounts are indices representing parameters related to the colors assigned to each pixel included in the handwritten numeric imagery.

[0023] 3 is a diagram showing an example of partial data according to an embodiment of the present invention. The partial data is data showing all feature quantities defined for a part of the samples included in the entire data, as shown in FIG. 3, and is expressed as a matrix. The entire data shown in FIG. 2 is, for example, the partial data X1, ... and partial data X c It can be divided into:

[0024] Furthermore, terminals 10-1, ... and terminal 10-c all possess first anchor data. The first anchor data is data that is commonly distributed to terminals 10-1, ... and terminal 10-c, and is data that does not indicate information that needs to be kept secret by the organizations that manage terminals 10-1, ... or terminal 10-c.

[0025] The first anchor data is the partial data X1, ..., and the partial data X c It is preferable that the first anchor data is data close to the partial data X1, ... and the partial data X c The data is close to the range of the values ​​included in the first anchor data, which can be taken by partial data X1, ..., and partial data X c It means that the numerical value included in is close to the range that can be taken. Alternatively, the first anchor data is partial data X1, ... and partial data X c The data is close to the attribute included in the first anchor data, which is the partial data X1, ..., and the partial data X c It means that the attribute is close to the attribute contained in

[0026] However, the first anchor data may be randomly generated data. For example, the first anchor data may be partial data X1, ..., and partial data X c Alternatively, the first anchor data may be data representing a random number in the range from the minimum value to the maximum value of the feature amount in at least one of the partial data X1, ..., or the partial data X cThe first anchor data may be data generated by applying a low-rank approximation to the first anchor data and then adding a perturbation to the first anchor data. The first anchor data may also be data generated by each of the terminals 10-1, ... and 10-c and exchanged between the terminals 10-1, ... and 10-c.

[0027] Furthermore, terminals 10-1, ... and terminal 10-c all possess second anchor data. The second anchor data is data that is commonly distributed to terminals 10-1, ... and terminal 10-c, and is data that does not indicate information that needs to be kept secret by the organizations that manage terminals 10-1, ... or terminal 10-c.

[0028] The second anchor data is the partial data X1, ..., and the partial data X c It is preferable that the second anchor data is data close to the partial data X1, ... and the partial data X c The data is close to the range of the values ​​included in the second anchor data, which can be taken by partial data X1, ..., and partial data X c It means that the numerical value included in is close to the range that can be taken. Alternatively, the second anchor data is partial data X1, ... and partial data X c The data is close to the attribute included in the second anchor data, and the attribute included in the second anchor data is the partial data X1, ..., and the partial data X c It means that the attribute is close to the attribute contained in

[0029] However, the second anchor data may be randomly generated data. For example, the second anchor data may be partial data X1, ..., and partial data X c Alternatively, the second anchor data may be data representing a random number in the range from the minimum value to the maximum value of the feature amount in at least one of the partial data X1, ..., or the partial data X c The second anchor data may be data generated by applying a low-rank approximation to the first anchor data and then adding a perturbation to the second anchor data. The second anchor data may also be data generated by each of the terminals 10-1, ... and 10-c and exchanged between the terminals 10-1, ... and 10-c.

[0030] The first anchor data and the second anchor data may be the same data or different data. Furthermore, terminal 10-1 holds partial data X1, the first anchor data, and the second anchor data in a storage medium that it can access, for example. This storage medium only needs to be accessible from terminal 10-1, and may or may not be included in terminal 10-1. Terminals 10-2, ..., terminal 10-c similarly store partial data X2, the first anchor data, the second anchor data, ..., partial data X in their storage media, respectively. c , first anchor data and second anchor data.

[0031] Next, an example of the functional configuration of the integrated analytical model generation system according to the embodiment will be described with reference to FIGS.

[0032] 4 is a diagram showing an example of the functional configuration of a terminal according to an embodiment of the present invention. As shown in FIG. 4, terminal 10-1 includes an abstraction function generation unit 11-1, an abstraction execution unit 12-1, an abstraction function discarding unit 13-1, a data transmission unit 14-1, a data reception unit 15-1, and a model generation unit 16-1.

[0033] Similarly, terminal 10-2, ..., terminal 10-c respectively include an abstraction function generation unit 11-2, an abstraction execution unit 12-2, an abstraction function discarding unit 13-2, a data transmission unit 14-2, a data receiving unit 15-2, and a model generation unit 16-2, and ..., an abstraction function generation unit 11-c, an abstraction execution unit 12-c, an abstraction function discarding unit 13-c, a data transmission unit 14-c, a data receiving unit 15-c, and a model generation unit 16-c.

[0034] 5 is a diagram showing an example of the functional configuration of a data processing device according to an embodiment of the present invention. As shown in FIG. 5, the data processing device 20 includes a data receiving unit 21, a label data generating unit 22, and a data transmitting unit 23.

[0035] The abstraction function generator 11-1 generates an abstraction function f1 used to perform abstraction processing on the partial data X1, the first anchor data, and the second anchor data. Similarly, the abstraction function generators 11-2, ..., and 11-c generate abstraction functions f2, ..., and f3, respectively. c Generate.

[0036] Furthermore, at least one of the abstraction function generation units 11-1, ... and 11-c may generate an abstraction function by at least one of a process of randomly extracting a portion of the partial data and a process of imparting a random element to the abstraction function. The process of randomly extracting a portion of the partial data is, for example, downsampling. The process of imparting a random element to the abstraction function is, for example, a process of imparting a random perturbation to the abstraction function.

[0037] Furthermore, at least one of the abstraction function generation units 11-1, ... and the abstraction function generation unit 11-c may randomly extract a portion of the partial data according to a rule determined by a predetermined method based on, for example, the time when the process of generating the abstraction function started. Also, at least one of the abstraction function generation units 11-1, ... and the abstraction function generation unit 11-c may impart a random perturbation to the abstraction function determined by a predetermined method based on, for example, the time when the process of generating the abstraction function started.

[0038] It is preferable that these times be measured in fine units such as microseconds, because if times measured in fine units are used, it will be difficult for the person managing terminal 10-1, ... or terminal 10-c to understand the relationship between time and randomness, making it difficult to generate the same abstraction function and difficult to determine the information indicated by the partial data from the abstract partial data, which will be described later.

[0039] The abstraction execution unit 12-1 applies an abstraction function f1 to the partial data X1 to generate abstracted partial data W1. Similarly, the abstraction execution units 12-2, ..., and 12-c apply an abstraction function f1 to the partial data X2, ..., and the partial data Xc abstract function f2, ... and abstract function f c Applying the above to abstract partial data W2, ... and abstract partial data W c Generate.

[0040] The abstracted partial data W1 is numerical data obtained by abstracting the feature quantities contained in the partial data X1 using a unique abstraction process represented by an abstraction function f1. Similarly, the abstracted partial data W2 is numerical data obtained by abstracting the feature quantities contained in the partial data X2 using a unique abstraction process represented by an abstraction function f2. In addition, the abstracted partial data W c is an abstract function f c The partial data X is generated by the original abstraction process c It is numerical data obtained by abstracting the features contained in

[0041] The abstraction execution unit 12-1 may equally and randomly interchange the rows of the abstraction partial data W1 and the label data Y1. Similarly, the abstraction execution units 12-2, ..., and 12-c may equally and randomly interchange the rows of the abstraction partial data W1, the label data Y2, ..., and the abstraction partial data W c and label data Y C The rows of may be interchanged equally.

[0042] The abstraction execution unit 12-1 applies the abstraction function f1 to the first anchor data to generate abstracted first anchor data. Similarly, the abstraction execution units 12-2, ..., and 12-c apply the abstraction functions f2, ..., and f3 to the first anchor data, ..., and the first anchor data, respectively. c is applied to generate abstracted first anchor data, ... and abstracted first anchor data.

[0043] The first abstract anchor data W generated by the terminal 10-1 anc1 1 is data representing a linear combination of feature quantities included in the first anchor data by a unique abstraction process represented by an abstraction function f1. Similarly, the first abstract anchor data W generated by the terminal 10-2anc1 2 is data representing a linear combination of the feature quantities included in the first anchor data by a unique abstraction process represented by the abstraction function f2. Also, the first abstract anchor data W generated by the terminal 10-c anc1 c is an abstract function f c This is data that represents a linear combination of the features included in the first anchor data by a unique abstraction process represented by:

[0044] Similarly, the abstraction execution unit 12-2, ..., and the abstraction execution unit 12-c apply the abstraction function f2, ..., and the abstraction function f3, ..., to the second anchor data, ..., and the second anchor data, respectively. c is applied to generate abstracted second anchor data, ... and abstracted second anchor data.

[0045] The second abstract anchor data W generated by the terminal 10-1 anc2 1 is data representing a linear combination of the feature quantities included in the second anchor data by a unique abstraction process represented by the abstraction function f1. Similarly, the second abstract anchor data W generated by the terminal 10-2 anc2 2 is data representing a linear combination of the feature quantities included in the second anchor data by a unique abstraction process represented by the abstraction function f2. Also, the second abstract anchor data W generated by the terminal 10-c anc2 c is an abstract function f c This data represents a linear combination of the features included in the second anchor data through a unique abstraction process represented by:

[0046] The above-mentioned abstraction process uses a linear or nonlinear dimensionality reduction method that uses an objective function. Dimensionality reduction methods are broadly divided into supervised and unsupervised dimensionality reduction methods. Examples of supervised dimensionality reduction methods include Linear Discriminant Analysis and Local Fisher Discriminant Analysis. Examples of unsupervised dimensionality reduction methods include t-distributed Stochastic Neighbor Embedding (t-SNE), Principal Component Analysis, Local Linear Embedding, Local Tangent Space Alignment, and Locality Preserving Projections. Alternatively, the above-mentioned abstraction process uses a substructure of deep learning.

[0047] The abstract function discarding unit 13-1 discards the abstracted partial data W1, the first abstracted anchor data W anc1 1 and the second abstract anchor data W anc2 After the abstract function discarding unit 13-1 generates the abstraction partial data W2, the abstraction anchor data W3, and the abstraction function discarding unit 13-2, ..., and the abstraction function discarding unit 13-c discard the abstraction partial data W3, the first abstraction anchor data W4, and the abstraction anchor data W5, respectively. anc1 2 and the second abstraction anchor data W anc2 2, ..., Abstraction partial data W c , the first abstract anchor data W anc1 c and second abstraction anchor data W anc2 c After the abstract functions f2, ... and f c Discard.

[0048] The abstract function discarding unit 13-1 also discards the abstracted partial data W1 and the first abstracted anchor data W anc1 1 and the second abstract anchor data W anc2It is preferable to discard the abstract function f1 before the abstract partial data W2, the first abstract anchor data W1 are transmitted to the data processing device 20. Similarly, the abstract function discarding units 13-2, ..., and 13-c discard the abstract partial data W2, the first abstract anchor data W1, and the anc1 2 and the second abstraction anchor data W anc2 2, ..., Abstraction partial data W c , the first abstract anchor data W anc1 c and second abstraction anchor data W anc2 c Before being transmitted to the data processing device 20, the abstraction functions f2, ... and the abstraction function f c It is preferable to discard the

[0049] The data transmission unit 14-1 receives the abstracted partial data W1, the first abstracted anchor data W anc1 1. Second Abstraction Anchor Data W anc2 1 and label data Y1 to the data processing device 20. Similarly, the data transmission units 14-2, ... and 14-c transmit the abstraction partial data W2, the first abstraction anchor data W anc1 2. Second Abstraction Anchor Data W anc2 2 and label data Y2, ..., abstract partial data W c , the first abstract anchor data W anc1 c , the second abstract anchor data W anc2 c and label data Y c is transmitted to the data processing device 20.

[0050] The data receiving unit 21 receives the abstraction partial data W1, the first abstraction anchor data W anc1 1. Second Abstraction Anchor Data W anc2 Similarly, the data receiving unit 21 receives the abstraction partial data W2, the first abstraction anchor data W anc1 2. Second Abstraction Anchor Data W anc2 c and label data Y2..., abstraction part data W c , the first abstract anchor data W anc1c , the second abstract anchor data W anc2 c and label data Y C are received from the terminals 10-2, . . . and 10-c, respectively.

[0051] The label data generating unit 22 generates the first abstract anchor data W received from the terminal 10-1. anc1 1, ... and the first abstract anchor data W received from the terminal 10-c anc1 c Abstract partial data W1, ... and abstract partial data W c The label data generating unit 22 generates the integrated data Z by integrating the first abstract anchor data W anc1 1... and first abstract anchor data W anc1 c Second abstraction anchor data using W anc2 1, ... and second abstract anchor data W anc2 c By integrating the second anchor data Z anc2 Generate.

[0052] Next, the label data generating unit 22 generates the label data of the elements Z1, . . . and Z c Label data Y1, ..., and label data Y c An integrated analysis model h is generated that outputs the above based on the integrated data Z.

[0053] Next, the label data generating unit 22 generates the integrated second anchor data Z anc2 and the integrated second anchor data Z from the integrated analysis model h anc2 Element Z contained in anc2 1, … and element Z anc2 c Second label data Y indicating the label corresponding to the content indicated by each anc2 1, ... and second label data Y anc2 c Generate the second label data Y anc2 1, ... and second label data Y anc2 care all integrated second anchor data Z anc2 This indicates the label corresponding to the content indicated by the element included in the tag.

[0054] Specifically, the label data generating unit 22 generates the first abstract anchor data W received from the terminal 10-1. anc1 Similarly, the label data generating unit 22 converts the first abstract anchor data W received from the terminal 10-2 into provisional integration data using the provisional integration function e1. anc1 Similarly, the label data generating unit 22 converts the first abstract anchor data W received from the terminal 10-c into provisional integration data using the provisional integration function e2. anc1 c The temporary integration function e c Convert the data into provisional integration data using the provisional integration function e1, ..., provisional integration function e c is a linear or non-linear function.

[0055] Next, the label data generating unit 22 calculates temporary integration functions e1, . . . , e2 that reduce the differences between these c pieces of temporary integration data. c The generation unit 220 calculates the temporary integration functions e1, ..., e2 that minimize the differences between these c pieces of temporary integration data. c integration function g1, ..., integration function g c It is preferable to calculate it as follows.

[0056] Integration function g1, ..., integration function g c The problem of calculating reduces to, for example, a minimization problem. In particular, when the provisional integration function is a linear function, this problem reduces to a generalized total least squares (TLS) problem. The label data generation unit 22 calculates the integration functions g1, ..., integration function g c Generate an integrated analytical model h based on the label data Y1, ... and label data Y cThe data Y obtained by combining the above data, the integrated data Z, and the integrated analytical model h have a relationship expressed as Y≈h(Z). anc2 and the second label data Y using the integrated analysis model h anc2 1, ... and second label data Y anc2 c Generate.

[0057] The data transmission unit 23 receives the second label data Y anc2 1 to the terminal 10-1. Similarly, the data transmitting unit 23 transmits the second label data Y anc2 2, ... and second label data Y anc2 c are transmitted to the terminals 10-2, . . . and 10-c, respectively.

[0058] The data receiving unit 15-1 receives the second label data Y anc2 1 from the data processing device 20. Similarly, the data receiving units 15-2, ... and 15-c each receive the second label data Y anc2 2, ... and second label data Y anc2 c is received from the data processing device 20.

[0059] The model generation unit 16-1 generates the second anchor data and the second label data Y anc2 1 and the content indicated by the second anchor data and the second label data Y anc2 Similarly, the model generators 16-2, ..., and 16-c generate integrated analytical models t2, ..., and t3, respectively. c The integrated analytical model t1 is a model used for the partial data newly received by the terminal 10-1. Similarly, the integrated analytical model t2, ... and the integrated analytical model t c are the models used for the partial data newly received by terminal 10-2, ..., and the partial data newly received by terminal 10-c, respectively.

[0060] Next, an example of processing executed by terminals 10-1, ..., terminal 10-c, and data processing device 20 will be described with reference to Fig. 6. Fig. 6 is a sequence diagram showing an example of processing executed by terminals and data processing device according to an embodiment of the present invention. In the description using Fig. 6, terminal 10-1 will be taken as an example, but the same applies to terminals 10-2, ..., terminal 10-c.

[0061] In step S10, the abstraction function generating unit 11-1 generates an abstraction function.

[0062] In step S20, the abstraction execution unit 12-1 generates abstraction partial data, first abstraction anchor data, and second abstraction anchor data.

[0063] In step S30, the abstract function discarding unit 13-1 discards the abstract function.

[0064] In step S40, the data transmitting unit 14-1 transmits the abstraction partial data, the first abstraction anchor data, the second abstraction anchor data, and the label data.

[0065] In step S50, the data receiving unit 21 receives the abstraction partial data, the first abstraction anchor data, the second abstraction anchor data, and the label data.

[0066] In step S60, the label data generating unit 22 generates integrated data and generates an integrated analysis model.

[0067] In step S70, the label data generating unit 22 generates integrated second anchor data, and generates second label data using the integrated second anchor data and the integrated analysis model.

[0068] In step S80, the data transmitting unit 23 transmits the second label data.

[0069] In step S90, the data receiving unit 15-1 receives the second label data.

[0070] In step S100, the model generation unit 16-1 generates an integrated analysis model that indicates the relationship between the content indicated by the second anchor data and the content indicated by the second label data.

[0071] The integrated analytical model generation system 1 according to the embodiment has been described above. The integrated analytical model generation system 1 includes terminals 10-1 and the like and a data processing device 20. The terminals 10-1 and the like generate an abstraction function, generate abstraction partial data, first abstraction anchor data, and second abstraction anchor data using the abstraction function, discard the abstraction function, and transmit these three pieces of data and label data to the data processing device 20. The data processing device 20 uses the first abstraction anchor data to integrate multiple abstraction partial data to generate integrated data Z, generates an integrated analytical model h that outputs label data indicating labels corresponding to content indicated by elements included in the integrated data Z based on the integrated data Z, integrates multiple second abstraction anchor data to generate integrated second anchor data, and generates second label data using the integrated second anchor data and the integrated analytical model h to transmit the second label data to the terminal 10-1 and the like. The terminals 10-1 and the like use the second anchor data and the second label data to generate an integrated analytical model indicating the relationship between the content indicated by the second anchor data and the content indicated by the second label data.

[0072] As a result, the integrated analysis model generation system 1 discards the abstraction function after using the abstraction function, making it impossible to match partial data with abstracted partial data.

[0073] Furthermore, the terminals 10-1 and the like discard the abstraction function before the abstraction partial data, the first abstraction anchor data, and the second abstraction anchor data are transmitted to the data processing device. This allows the integrated analysis model generation system 1 to more reliably prevent a situation in which, after the abstraction function is used, the abstraction function is leaked to someone other than the person managing the terminals 10-1 and the like, resulting in partial data and abstracted partial data being compared.

[0074] Furthermore, the terminals 10-1 and the like generate an abstraction function by at least one of a process of randomly extracting a portion of the partial data and a process of adding a random element to the abstraction function, which makes it impossible for the integrated analysis model generation system 1 to generate an abstraction function identical to the abstraction function in question, thereby more reliably preventing partial data from being matched with abstracted partial data.

[0075] Furthermore, the terminals 10-1 and the like randomly interchange the rows of the abstracted partial data and the label data, thereby making it difficult for the integrated analysis model generation system 1 to match the partial data with the abstracted partial data, and more reliably preventing partial data from being matched with the abstracted partial data.

[0076] Next, with reference to FIGS. 7 and 8, a result of applying the integrated analytical model generation system according to the embodiment to a specific example will be described.

[0077] 7 is a diagram showing an example of the relationship between the number of institutions and the accuracy rate when abstracted partial data indicating corporate ratings are integrated and analyzed using an integrated analysis model generated by a terminal according to an embodiment of the present invention. The overall data in the case shown in Fig. 7 consists of 3,932 pieces of data having feature quantities of working capital / total assets, retained earnings / total assets, earnings before interest and taxes / total assets, market capitalization / book value of all debt, sales / total assets, and industry, and seven labels from AAA to CCC are assigned to each piece of data.

[0078] The first anchor data in the case shown in Figure 7 is data representing uniform random numbers ranging from the minimum value to the maximum value of each feature amount. The second anchor data in this case is data generated by performing low-rank approximation using singular value decomposition on partial data held by each institution, adding noise, and linearly combining the data.

[0079] The abstraction function shown in Figure 7 is an abstraction function generated by applying principal component analysis to data with random perturbations for each institution. Note that because principal component analysis is a dimensionality reduction method that depends on the partial data, the abstraction method will differ for each institution. Also, in the case shown in Figure 7, the rows of the abstracted partial data are randomly permuted.

[0080] The integration function in the case shown in Figure 7 is generated by solving a generalized total least squares problem. The integrated data in the case shown in Figure 7 is analyzed using kernelized ridge regression with a Gaussian kernel. The integrated analysis model generated by the device in the case shown in Figure 7 is generated using kernelized ridge regression with a Gaussian kernel.

[0081] The horizontal axis of FIG. 7 represents the number of institutions. The vertical axis of FIG. 7 represents the accuracy rate of the analysis results. The solid line in FIG. 7 represents the relationship between the number of institutions and the accuracy rate when the abstracted partial data indicating corporate ratings is integrated and analyzed using the integrated analysis model generated by the terminal according to the embodiment. The dashed line in FIG. 7 represents the relationship between the number of institutions and the accuracy rate when the partial data indicating corporate ratings is analyzed without performing processing to prevent matching between the abstracted partial data and the partial data. The dashed-dotted line in FIG. 7 represents the relationship between the number of institutions and the accuracy rate when the partial data indicating corporate ratings is integrated and analyzed instead of the abstracted partial data. The dashed-dotted line in FIG. 7 represents the relationship between the number of institutions and the accuracy rate when one piece of partial data is analyzed only by the institution that holds that partial data.

[0082] Comparing the solid line shown in Figure 7 with the dashed line shown in Figure 7, it can be seen that the accuracy rate when the abstracted partial data indicating the ratings of companies is integrated and analyzed using the integrated analysis model generated by the terminal of the embodiment is lower than the accuracy rate when the analysis is performed without performing processing to make it impossible to match with the abstracted partial data, but still shows a relatively high accuracy rate.

[0083] Furthermore, when comparing the solid line shown in Figure 7 with the dotted line shown in Figure 7, it can be seen that the accuracy rate when the abstracted partial data indicating the ratings of companies is integrated and analyzed using the integrated analysis model generated by the terminal of the embodiment is higher than the accuracy rate when the partial data is integrated and analyzed.

[0084] Furthermore, when comparing the solid line shown in Figure 7 with the two-dot chain line shown in Figure 7, it can be seen that the accuracy rate when abstracted partial data indicating corporate ratings are integrated and analyzed using the integrated analysis model generated by the terminal of the embodiment is higher than the accuracy rate when one piece of partial data is analyzed only by the institution that holds that partial data.

[0085] 8 is a diagram showing an example of the relationship between the number of institutions and the accuracy rate when abstracted partial data representing handwritten numbers are integrated and analyzed using an integrated analysis model generated by a terminal according to an embodiment of the present invention. The horizontal axis of FIG. 8 represents the number of institutions.

[0086] The overall data in the case shown in Fig. 8 is data representing handwritten numbers from "0" to "9" displayed as a grayscale image with 28 vertical and horizontal pixels. Other conditions, procedures, etc. in the case shown in Fig. 8 are the same as those in the case shown in Fig. 7.

[0087] The vertical axis in Figure 8 represents the accuracy rate of the analysis results. The solid line in Figure 8 represents the relationship between the number of institutions and the accuracy rate when abstracted partial data representing handwritten digits are integrated and analyzed using the integrated analysis model generated by the terminal according to the embodiment. The dashed line in Figure 8 represents the relationship between the number of institutions and the accuracy rate when the partial data representing handwritten digits is analyzed without performing processing to prevent matching between the partial data representing handwritten digits and the abstracted partial data. The dashed-dotted line in Figure 8 represents the relationship between the number of institutions and the accuracy rate when partial data representing handwritten digits is integrated and analyzed, rather than the abstracted partial data. The dashed-two-dot line in Figure 8 represents the relationship between the number of institutions and the accuracy rate when one piece of partial data is analyzed only by the institution that holds that partial data.

[0088] Comparing the solid line shown in Figure 8 with the dashed line shown in Figure 8, it can be seen that the accuracy rate when abstracted partial data representing handwritten numbers are integrated and analyzed using the integrated analysis model generated by the terminal of the embodiment is higher than the accuracy rate when the analysis is performed without performing processing to make it impossible to match the abstracted partial data.

[0089] Furthermore, when comparing the solid line shown in Figure 8 with the dotted line shown in Figure 8, it can be seen that the accuracy rate when the abstract partial data representing handwritten numbers is integrated and analyzed using the integrated analysis model generated by the terminal of the embodiment is slightly lower than the accuracy rate when the partial data is integrated and analyzed.

[0090] Furthermore, when comparing the solid line shown in Figure 8 with the two-dot chain line shown in Figure 8, it can be seen that the accuracy rate when abstracted partial data representing handwritten numbers are integrated and analyzed using the integrated analysis model generated by the terminal of the embodiment is higher than the accuracy rate when one piece of partial data is analyzed only by the institution that holds that partial data.

[0091] At least some of the functions of terminals 10-1, ..., terminal 10-c, and data processing device 20 may be realized by hardware including circuitry executing a program. The hardware referred to here is, for example, a CPU (Central Processing Unit), an LSI (Large Scale Integration), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a GPU (Graphics Processing Unit). The above-mentioned program is stored in a storage device having a storage medium. The storage medium referred to here is, for example, a HDD (Hard Disk Drive), a flash memory, a ROM (Read Only Memory), or a DVD (Digital Versatile Disc). Furthermore, the above-mentioned program may be a differential program that realizes some of the functions of terminals 10-1, ..., terminal 10-c, and data processing device 20.

[0092] The above-described integrated analytical model generation system 1 can also be applied to data other than data related to a company's financial affairs and data related to handwritten number images. For example, the above-described integrated analytical model generation system 1 can also be applied to data related to agriculture, data related to smart cities, and data related to development held by companies in the manufacturing industry.

[0093] The embodiments of the present invention have been described above with reference to the drawings. However, the integrated analytical model generation system 1 is not limited to the above-described embodiments, and various modifications, substitutions, combinations, and design changes can be made without departing from the spirit of the present invention. [Explanation of symbols]

[0094] 1...integrated analysis model generation system, 10-1,...,10-c...terminal, 11-1,...,11-c...abstraction function generation unit, 12-1,...,12-c...abstraction execution unit, 13-1,...,13-c...abstraction function discarding unit, 14-1,...,14-c...data transmission unit, 15-1,...,15-c...data receiving unit, 16-1,...,16-c...model generation unit, 20...data processing device, 21...data receiving unit, 22...label data generation unit, 23...data transmission unit

Claims

1. an abstraction function generation unit that generates an abstraction function used to perform abstraction processing on the partial data, the first anchor data, and the second anchor data; an abstraction execution unit that applies the abstraction function to the partial data to generate abstracted partial data, applies the abstraction function to the first anchor data to generate first abstracted anchor data, and applies the abstraction function to the second anchor data to generate second abstracted anchor data; an abstraction function discarding unit that discards the abstraction function after the abstraction partial data, the first abstraction anchor data, and the second abstraction anchor data are generated; a data transmitting unit that transmits the abstraction partial data, the first abstraction anchor data, the second abstraction anchor data, and label data to a data processing device; a data receiving unit that receives, from the data processing device, second label data that indicates a label corresponding to content indicated by an element included in integrated second anchor data that is generated by integrating a plurality of second abstract anchor data using a plurality of first abstract anchor data; a model generation unit that uses the second anchor data and the second label data to generate an integrated analysis model that indicates a relationship between content indicated by the second anchor data and content indicated by the second label data; A terminal comprising:

2. the abstraction function discarding unit discards the abstraction function before the abstraction portion data, the first abstraction anchor data, and the second abstraction anchor data are transmitted to the data processing device. The terminal according to claim 1 .

3. the abstraction function generation unit generates the abstraction function by at least one of a process of randomly extracting a part of the partial data and a process of adding a random element to the abstraction function. The terminal according to claim 1 or claim 2.

4. the abstraction execution unit equally and randomly permutes rows of the abstract partial data and the label data; A terminal according to any one of claims 1 to 3.

5. a data receiving unit that receives abstraction partial data, first abstraction anchor data, second abstraction anchor data, and label data for each terminal; a label data generation unit that generates integrated data by integrating a plurality of the abstract partial data using the first abstraction anchor data received for each of the terminals, generates integrated second anchor data by integrating a plurality of the second abstraction anchor data using the first abstraction anchor data received for each of the terminals, generates an integrated analysis model that outputs label data indicating labels corresponding to content indicated by elements included in the integrated data based on the integrated data, and generates second label data indicating labels corresponding to content indicated by elements included in the integrated second anchor data using the integrated second anchor data and the integrated analysis model; a data transmission unit that transmits the second label data to the terminal; A data processing device comprising:

6. an abstraction function generation function that generates an abstraction function used to perform abstraction processing on the partial data, the first anchor data, and the second anchor data; an abstraction execution function that applies the abstraction function to the partial data to generate abstract partial data, applies the abstraction function to the first anchor data to generate first abstract anchor data, and applies the abstraction function to the second anchor data to generate second abstract anchor data; an abstraction function discarding function that discards the abstraction function after the abstraction partial data, the first abstraction anchor data, and the second abstraction anchor data are generated; a data transmission function for transmitting the abstraction portion data, the first abstraction anchor data, the second abstraction anchor data, and label data to a data processing device; a data receiving function that receives, from the data processing device, second label data that indicates labels corresponding to contents indicated by elements included in integrated second anchor data that is generated by integrating a plurality of the second abstract anchor data using a plurality of the first abstract anchor data; a model generation function that uses the second anchor data and the second label data to generate an integrated analysis model that indicates the relationship between the content indicated by the second anchor data and the content indicated by the second label data; A terminal control program that causes the terminal to execute the above.

7. a data receiving function for receiving the abstraction part data, the first abstraction anchor data, the second abstraction anchor data, and the label data for each terminal; a label data generation function that generates integrated data by integrating a plurality of the abstract partial data using the first abstraction anchor data received for each of the terminals, generates integrated second anchor data by integrating a plurality of the second abstraction anchor data using the first abstraction anchor data received for each of the terminals, generates an integrated analysis model that outputs label data indicating labels corresponding to content indicated by elements included in the integrated data based on the integrated data, and generates second label data indicating labels corresponding to content indicated by elements included in the integrated second anchor data using the integrated second anchor data and the integrated analysis model; a data transmission function for transmitting the second label data to the terminal; A data processing program that causes a data processing device to execute the above.

8. A terminal control method executed by a computer, comprising: the computer generates, by an abstraction function generation unit or an abstraction function generation function, an abstraction function used to perform abstraction processing on the partial data, the first anchor data, and the second anchor data; the computer, by an abstraction execution unit or an abstraction execution function, applies the abstraction function to the partial data to generate abstract partial data, applies the abstraction function to the first anchor data to generate first abstract anchor data, and applies the abstraction function to the second anchor data to generate second abstract anchor data; the computer discards the abstraction function after the abstraction partial data, the first abstraction anchor data, and the second abstraction anchor data are generated by an abstraction function discarding unit or an abstraction function discarding function; the computer transmits the abstraction partial data, the first abstraction anchor data, the second abstraction anchor data, and label data to a data processing device by a data transmission unit or a data transmission function; the computer receives, via a data receiving unit or a data receiver, second label data from the data processing device, which indicates labels corresponding to content indicated by elements included in integrated second anchor data generated by integrating a plurality of the second abstract anchor data using a plurality of the first abstract anchor data; the computer generates, by a model generation unit or a model generation function, an integrated analysis model that indicates a relationship between the content indicated by the second anchor data and the content indicated by the second label data, using the second anchor data and the second label data; Terminal control methods.

9. A data processing method executed by a computer, comprising: the computer receives abstraction partial data, first abstraction anchor data, second abstraction anchor data, and label data for each terminal by a data receiving unit or a data receiving function; the computer, by a label data generation unit or a label data generation function, generates integrated data by integrating a plurality of the abstract partial data using the first abstraction anchor data received for each of the terminals, generates integrated second anchor data by integrating a plurality of the second abstraction anchor data using the first abstraction anchor data received for each of the terminals, generates an integrated analysis model that outputs label data indicating labels corresponding to content indicated by elements included in the integrated data based on the integrated data, and generates second label data indicating labels corresponding to content indicated by elements included in the integrated second anchor data using the integrated second anchor data and the integrated analysis model; the computer transmits the second label data to the terminal by a data transmission unit or a data transmission function; Data processing methods.

Images