JP2017126112A - Server, distributed server system, and information processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To protect personal information, while collecting information to be used in statistic/machine learning.SOLUTION: A server includes: an existing data storage unit 21 which stores data on an existing user; a data acquisition unit 22 which acquires data on a user different from the existing user; a difference calculation unit 23 which calculates difference data between the data on the existing user stored in the existing data storage unit 21 and the data on another user acquired by the data acquisition unit 22; a difference data storage unit 24 which stores the difference data calculated by the difference calculation unit 23; and a statistic unit 25 which collects statistics on users by use of the existing-user data and the difference data.SELECTED DRAWING: Figure 3

Description

  The present invention relates to a server, a distributed server system, and an information processing method.

  Conventionally, when a customer's personal information is used for statistics and machine learning, a system that asks the customer for consent to use the personal information is already known. On the other hand, there are many users who are reluctant to provide personal information, no matter how much consent is obtained.

  On the other hand, for example, Patent Document 1 uses a privacy protection collaborative filtering technique to encrypt an evaluation value for each item provided by the user, or to assign a random number to the evaluation value for each item provided. It is disclosed that it is difficult to obtain the original evaluation value so that personal information is not directly passed to the user.

  However, for example, data specialized for a specific method in the prior art such as Patent Document 1 cannot collect data used for general statistical / machine learning.

  The present invention has been made in view of the above-described circumstances, and an object thereof is to protect personal information while collecting information that can be used in statistics and machine learning.

  In order to solve the above-described problems, the server of the present invention includes an existing data storage unit that stores data related to an existing user, a data acquisition unit that acquires data related to a user different from the existing user, and an existing data storage unit. Difference calculation means for calculating difference data between data relating to an existing user stored in the data and data relating to another user acquired by the data acquisition means, and difference data for storing difference data calculated by the difference calculation means Storage means, and statistical means for collecting statistics on user data using data and difference data on existing users are provided.

  According to the present invention, personal information can be protected while collecting information that can be used in statistics and machine learning.

It is a schematic block diagram of the system containing the server in embodiment of this invention. It is a hardware block diagram of the server in embodiment of this invention. It is a functional block diagram of a server in an embodiment of the present invention. It is an image figure which shows the process example in embodiment of this invention. It is a flowchart which shows the process sequence in embodiment of this invention. It is a functional block diagram of a distributed server system in an embodiment of the present invention. It is a functional block diagram of a distributed server system in an embodiment of the present invention. It is a functional block diagram of a distributed server system in an embodiment of the present invention. It is a functional block diagram of a distributed server system in an embodiment of the present invention.

  A server and a distributed server system according to an embodiment of the present invention will be described below with reference to the drawings. In addition, in each figure, the same code | symbol is attached | subjected to the part which is the same or it corresponds, The duplication description is simplified thru | or abbreviate | omitted suitably. The embodiment described below is the best mode of the present invention, and does not limit the scope of the claims according to the present invention.

  In the present embodiment, the “server” refers to a program for providing some service to a client in a so-called client server system. On the other hand, it goes without saying that the “server” includes the meaning of an information processing apparatus that operates according to the program.

<About the schematic configuration of a system including a server in this embodiment>
A schematic configuration of a system including a server in the present embodiment will be described with reference to FIG. The system 1 in this embodiment is configured by connecting a server 10 and a client 20 via a network 30. In FIG. 1, the server 10 and the client 20 are one, but a configuration in which a plurality of servers and a plurality of clients are connected via a network may be used.

  The server 10 functions as an information processing apparatus such as a personal computer (hereinafter referred to as “PC”), a storage or a hard disk having an information processing function. Moreover, as long as it has the processing capability which implement | achieves a server function, it is not limited to PC etc.

  The client 20 corresponds to an information processing apparatus such as a PC used by a user who uses the server 10 and a portable information terminal such as a smartphone or a tablet.

  The communication form of the network 30 may be a LAN (Local Area Network) or a WAN (Wide Area Network). The network 30 may be a wireless communication network or a wired communication network.

<About the hardware configuration of the server in this embodiment>
The hardware configuration of the server 10 in this embodiment will be described with reference to FIG. The server 10 in the present embodiment includes a CPU 11, a RAM 12, a ROM 13, an NW I / F 14, an HDD 15, an input unit 16, and an output unit 17 as hardware configurations. Note that these are examples of configurations for the server 10 to execute functions (processes) described later, and hardware other than these is not excluded.

  The CPU 11 is a main control unit for realizing each process to be described later in the server 10. The CPU 11 implements each function in the server 10 by executing a program in which a processing program defining each processing stored in the ROM 13 is read into the RAM 12.

  The RAM 12 is a storage unit that functions as a work memory of the CPU 11 as described above. The ROM 13 is a storage unit that stores a processing program that defines each process as described above, various parameters required for control of the server 10, and the like.

  The NW I / F 14 is a network interface for connecting the server 10 and the client 20 to each other via the network 30 shown in FIG. The HDD 15 is a large-capacity storage unit that stores, for example, data acquired from the client 20.

  The input unit 16 is an input device such as a keyboard or a mouse. Moreover, the device which receives a user's touch operation like the touch panel superimposed on the display part mentioned later may be sufficient. Furthermore, the input unit 16 may include a camera that acquires an image by capturing a video and a microphone that receives audio input.

  The output unit 17 is a display unit such as a display. The output unit 17 may include a speaker that outputs sound.

<Regarding Server Functional Blocks in the Embodiment of the Present Invention>
Functional blocks of the server 10 in the embodiment of the present invention will be described with reference to FIG. The server 10 in the present embodiment includes an existing data storage unit 21, a data acquisition unit 22, a difference calculation unit 23, a difference data storage unit 24, a statistics unit 25, and a display unit 26 as functional blocks.

  The existing data storage unit 21 is an existing data storage unit that stores data relating to an existing user, and the function is realized by the CPU 11, the ROM 13, and the HDD 15 shown in FIG. In this embodiment, the “data related to the user” is personal information (hereinafter referred to as “patient data”) provided by the patient mainly in the medical field such as the user's name, age, height, weight, blood pressure, medical history, and the like. This will be described as an example. On the other hand, the data regarding the user is not limited to this example. Further, “existing” means that it is already stored in a storage device such as the HDD 15.

  The data acquisition unit 22 is a data acquisition unit that acquires data related to a user other than the existing user, and the function is realized by the CPU 11, the RAM 12, and the NW I / F 14 illustrated in FIG. Although the data regarding another user to acquire is also said patient data, it is the data etc. which are not memorize | stored in the existing data storage part 21, for example, are acquired from the user different from the existing user. Note that patient data or the like does not necessarily have to go through a network. For example, you may make it acquire from a user via the input part 16 shown in FIG. Further, it may be acquired by reading data stored in an external storage medium.

  The difference calculation unit 23 is a difference calculation unit that calculates difference data between data related to an existing user stored in the existing data storage unit 21 and data related to another user acquired by the data acquisition unit 22. The function of the difference calculation unit 23 is realized by the CPU 11 and the RAM 12 illustrated in FIG.

  In the present embodiment, the difference calculation unit 23 calculates difference data between the average value of data related to a plurality of existing users stored in the existing data storage unit 21 and the data related to the acquired other users. To do. For example, even in the difference data between the average value of the user of the same age and the user in the data regarding the acquired another user, the difference data between the average value of the user having the same height or the same weight in the data regarding the acquired another user But you can.

  The difference data storage unit 24 is a difference data storage unit that stores the difference data calculated by the difference calculation unit 23, and its function is realized by the CPU 11, the ROM 13, and the HDD 15 shown in FIG. Further, the difference data storage unit 24 may store difference data in which a difference from data relating to an existing user is greater than a predetermined threshold.

  Moreover, the difference data memory | storage part 24 should just select only the required difference data among the following, when not memorize | storing the difference data of all the items. For example, [1] an item with a large absolute value of difference, [2] [1] is an item with a large value obtained by dividing the difference by the standard deviation of the group, and [3] [2] It is only necessary to select a higher number of items among items having a large value by multiplying a value related to importance such as a contribution rate derived from. As in [2], by dividing the difference by the standard deviation of the group, it can be handled equally whether the original range of the numerical value is wide or narrow. Further, as in [3], by multiplying the contribution rate, it is possible to obtain information by selecting an item important for machine learning.

  Further, the difference data storage unit 24 may store the difference by narrowing down to N items, for example, as items having the largest difference. Here, in order to compare the magnitude of the difference, for example, a Mahalanobis distance, a distance obtained from the definition of the norm, or the like may be used. N may not be determined in advance, and the difference may be left by narrowing down to items having a certain difference or more. With the pattern recognition method that defines the distance, only items with a larger distance can be left.

  Further, the difference data storage unit 24 may select difference data depending on how many items have similar values with a threshold of several percent of the standard deviation in determining whether they are close. Further, the difference data storage unit 24 may store an item with a high contribution rate in statistics and machine learning, or a product obtained by multiplying a contribution rate by adding a standard deviation to the contribution rate. Thereby, only the user with unusual data can be taken in the pattern recognition method which defines distance.

  Furthermore, the difference data storage unit 24 may use not the difference between specific items but the difference between components obtained by principal component analysis for each item. At this time, the difference may be left as N components as components having a high contribution rate. Further, random noise may be added. By converting the data into data that is significantly different from the original data and acquiring the data, it becomes easier to convince the data, and it is possible to acquire items with lower importance in the principal component analysis.

  The statistical unit 25 is a statistical unit that collects statistics on the data on the user using the existing data on the user and the difference data, and the function is realized by the CPU 11, the RAM 12, the ROM 13, and the HDD 15 shown in FIG. The statistics unit 25 uses only necessary portions of the difference data stored in the difference data storage unit 24 when collecting statistics on data related to the user. Thereby, the processing load concerning the server 10 can be reduced.

  The “statistics” process by the statistics unit 25 includes an average value calculation process using data relating to an existing user and difference data. In the present embodiment, a value obtained by adding the value of the difference data to the value obtained by multiplying the existing data reference count by the item value is used.

  The “statistics” process by the statistical unit 25 includes a variance calculation process, whereby the original data is restored based on the difference data for only the item desired to be calculated. More specifically, the original data is restored by adding the difference and the average value. Further, in the “cross tabulation” processing by the statistics unit 25, the original data is restored based on the difference data for a plurality of items to be tabulated. Further, in the “feature amount calculation” process by the statistical unit 25, the original data restored after calculating the feature amounts one by one is discarded. With these processes, it is possible to perform statistical processing and learning processing by restoring the original data to the minimum.

  When the difference data calculated by the difference calculation unit 23 is stored in the difference data storage unit 24, the display unit 26 is a display unit that displays a permission confirmation screen for obtaining user permission for the storage of the difference data. . For example, when a user withdraws from a certain member organization, there is a usage form such as asking for consent to store all of the difference data, and displaying this screen when rejected. In addition, it cannot be overemphasized that the display part 26 can display the process result processed by the server 10 other than that.

<Example of processing in this embodiment>
A processing example in the present embodiment will be described with reference to FIG. Here, as the existing data group 101, “ID: 1, name: Ichiro Tanaka, height: 180, weight: 75, age: 38, blood pressure: 130, disease: none”, “ID: 2, name: Jiro Suzuki, Height: 174, Weight: 65, Age: 55, Blood pressure: 145, Disease: Diabetes, "ID: 1001, Name: 40's tall, Height: 175, Weight: 65, Age: 45, Blood pressure: 128, Disease : None "is displayed as an example.

  Further, “name: Saburo Sato, height: 173, weight: 55, age: 49, blood pressure: 130, illness: none” is displayed as the data 201 relating to another acquired user.

  Further, the difference data 301 between the data 201 relating to another user and the existing data indicated by ID: 1001 in the existing data group 101 is “name: xxx, height: −2, weight: −10, age: +4, blood pressure: +2, disease: none ”is displayed.

  Furthermore, the permission confirmation screen 401 shows that “Your weight is −10 and age is +4 compared to the average height information in your 40s. Can you provide this difference as data?” A screen with a confirmation message is displayed. As described above, the user information is stored as a difference, and when the difference is stored, the user information is required to be collected, so that the personal information can be protected while collecting information usable in statistics and machine learning. it can.

  Table 1 shows an example of an existing database stored in the existing data storage unit 21, and corresponds to data included in the existing data group 101 shown in FIG. That is, Table 1 shows data of Ichiro Tanaka with ID x1, and data of Jiro Suzuki with ID x2. In addition, the reference count of each data is stored. By putting the number of references in the database, the average value can be calculated without restoring the original data.

  Table 2 shows an example of the difference database stored in the difference data storage unit 24. The difference database includes “ID: y1, reference ID: x2, height: +5, weight +10, age +4, blood pressure −, illness: none”, “ID: y2, reference ID: x2, height: +10, weight − 10,-, blood pressure +10, illness: none ". Here, the difference between the data relating to two different users and the data of Jiro Suzuki in the existing database is shown.

<About the processing procedure in this embodiment>
A processing procedure in the present embodiment will be described with reference to FIG. First, a processing procedure related to calculation / storage of difference data will be described with reference to FIG. As a premise, it is assumed that data such as personal information acquired from the user so far is stored in the existing data storage unit 21 as existing data.

  First, the data acquisition unit 22 acquires data related to another user (step S1). Next, the difference calculation unit 23 searches for data similar to data related to another user from data related to an existing user stored in the existing data storage unit 21 (step S2).

  The difference calculating unit 23 calculates difference data between the searched data relating to the existing user and data relating to another user (step S3). The difference data storage unit 24 selects only a necessary difference from the calculated difference data (step S4).

  When only the necessary difference is selected, the display unit 26 displays the permission confirmation screen shown in FIG. 4 as to whether or not the difference data may be stored for the user, and asks the user for permission to store the difference data ( Step S5). When permission is obtained from the user, the difference data storage unit 24 stores the difference data (step S6).

  Next, a processing procedure relating to statistical processing in the present embodiment will be described with reference to FIG. First, the statistical unit 25 receives a request for statistical processing based on existing data and difference data from a user operating the server 10 (step S11).

  The statistical unit 25 restores only the minimum data to the original data based on the existing data and the difference data (step S12), and performs statistical processing using the restored original data (step S13). Thereafter, the statistical unit 25 discards the restored original data after completing the statistical processing (step S14).

<Regarding Functional Blocks of Distributed Server System in This Embodiment>
Functional blocks of the distributed server system A in this embodiment will be described with reference to FIG. In addition, about each functional block, description is abbreviate | omitted about the content which overlaps with the content demonstrated using FIG.

  FIG. 6 shows a configuration in which the server 500 and the server 600 are connected via a network. The server 500 has an existing data storage unit 21 and a difference data storage unit 24, and the server 600 has a data acquisition unit 22, a difference calculation unit 23, a statistics unit 25, and a display unit 26.

  In the distributed server system A, the processing is distributed so that the processing for storing data in the server 500 is main, and the processing from data acquisition to calculation processing is main in the server 600. As a result, data is collected and stored in the server 500, and the actual processing is performed by the server 600. Therefore, it is possible to reduce the processing load of the server 600 that performs batch management of data and actual processing.

  Next, functional blocks of the distributed server system B in the present embodiment will be described with reference to FIG. In addition, about each functional block, description is abbreviate | omitted about the content which overlaps with the content demonstrated using FIG.3 and FIG.6.

  FIG. 7 shows a configuration in which the server 500, the server 700, and the server 800 are connected via a network. The server 700 has a data acquisition unit 22 and a difference calculation unit 23, and the server 800 has a statistics unit 25 and a display unit 26.

  In the distributed server system B, processing is distributed so that the server 700 mainly performs data acquisition processing and difference calculation processing, and the server 800 mainly performs statistical processing and display processing. By distributing statistical processing and display processing, the server 700 can smoothly perform the difference calculation processing without load.

  Next, functional blocks of the distributed server system C in this embodiment will be described with reference to FIG. In addition, about each functional block, description is abbreviate | omitted about the content which overlaps with the content demonstrated using FIG.

  FIG. 8 shows a configuration in which the server 900 and the server 1000 are connected via a network. The server 900 has an existing data storage unit 21, and the server 1000 has a data acquisition unit 22, a difference calculation unit 23, a difference data storage unit 24, a statistics unit 25, and a display unit 26.

  In the distributed server system C, the server 900 performs processing for storing existing data as the main, and the server 1000 performs processing for other processing. Since the existing data is enormous and personal information, it is preferably managed by an administrator who has authority with a single server, and the distributed server system C is useful for such usage.

  Next, functional blocks of the distributed server system D in the present embodiment will be described with reference to FIG. In addition, about each functional block, description is abbreviate | omitted about the content which overlaps with the content demonstrated using FIG.3 and FIG.8.

  In the distributed server system D, processing is distributed so that the server 1100 mainly performs data acquisition processing, the server 1200 mainly performs difference calculation processing, and the server 1300 mainly stores difference data. . Further, the processing is distributed so that the server 1400 mainly performs statistical processing and the server 1500 mainly performs display processing.

  As in the distributed server system D, by causing each server to perform each process, it is possible to further reduce the processing load on each server.

  Note that the configuration of the distributed server systems A to D described above is merely an example, and it is needless to say that configurations other than these can be adopted. For example, the configuration is made according to the processing specifications of the server itself, the installation location of the server, and the like. Various changes can be made to the form.

  As described above, according to the present embodiment, the difference between the data acquired from a user different from the existing user and the data related to the existing user is stored, and the original data is calculated from the difference data and used for statistical processing. Therefore, the personal information of the user is not directly stored, which contributes to personal information protection. In addition, since the user is permitted to store the difference data, the personal information protection function is further strengthened. Further, the difference data itself is data that can be used for statistics and machine learning.

  Each of the above-described embodiments is a preferred embodiment of the present invention, and various modifications can be made without departing from the scope of the present invention. For example, each process in the server and distributed server system of the present embodiment described above can be executed using hardware, software, or a combined configuration of both.

  In the case of executing processing using software, it is possible to install and execute a program in which a processing sequence is recorded in a memory in a computer incorporated in dedicated hardware. Alternatively, the program can be installed and executed on a general-purpose computer capable of executing various processes.

1 System 10, 500, 600, 700, 800, 900, 1000, 1100, 1200, 1300, 1400, 1500 Server 20 Client 30 Network 11 CPU
12 RAM
13 ROM
14 NW I / F
15 HDD
16 Input unit 17 Output unit 21 Existing data storage unit 22 Data acquisition unit 23 Difference calculation unit 24 Difference data storage unit 25 Statistics unit 26 Display unit

JP 2014-016872 A

Claims (7)

Existing data storage means for storing data relating to existing users;
Data acquisition means for acquiring data relating to a user different from the existing user;
Difference calculation means for calculating difference data between data relating to an existing user stored in the existing data storage means and data relating to another user acquired by the data acquisition means;
Difference data storage means for storing difference data calculated by the difference calculation means;
Statistical means for taking statistics of data relating to the user using the data relating to the existing user and the difference data;
A server comprising:   2. The server according to claim 1, wherein the statistical unit uses only a necessary portion of the differential data stored in the differential data storage unit when collecting statistics on data relating to the user.   The server according to claim 1, wherein the difference data storage unit stores difference data having a difference from data relating to the existing user larger than a predetermined threshold.   When the difference data calculated by the difference calculation unit is stored in the difference data storage unit, the difference data storage unit includes a display unit that displays a permission confirmation screen for obtaining a user permission for storing the difference data. The server according to any one of claims 1 to 3.   The difference calculation means calculates difference data between an average value of data related to a plurality of existing users stored in the existing data storage means and data related to another user acquired by the data acquisition means. The server according to any one of claims 1 to 4, characterized in that: A distributed server system in which functions are distributed and stored in two or more storages,
Existing data storage means for storing data relating to existing users;
Data acquisition means for acquiring data relating to a user different from the existing user;
Difference calculation means for calculating difference data between data relating to an existing user stored in the existing data storage means and data relating to another user acquired by the data acquisition means;
Difference data storage means for storing difference data calculated by the difference calculation means;
Statistical means for taking statistics of data relating to the user using the data relating to the existing user and the difference data;
A distributed server system comprising: A computer storing data relating to an existing user in a storage unit;
A computer obtaining data relating to a user different from the existing user;
Calculating a difference data between the data related to the existing user stored in the storage unit and the acquired data related to the other user;
A computer storing the calculated difference data in the storage unit;
A computer taking statistics of data about the user using the data about the existing user and the difference data;
An information processing method comprising:

Images