WO2019137049A1 - 基于信息共享的预测方法、装置、电子设备及计算机存储介质 - Google Patents

基于信息共享的预测方法、装置、电子设备及计算机存储介质 Download PDF

Info

Publication number
WO2019137049A1
WO2019137049A1 PCT/CN2018/109728 CN2018109728W WO2019137049A1 WO 2019137049 A1 WO2019137049 A1 WO 2019137049A1 CN 2018109728 W CN2018109728 W CN 2018109728W WO 2019137049 A1 WO2019137049 A1 WO 2019137049A1
Authority
WO
WIPO (PCT)
Prior art keywords
real number
number vector
data
dimensional
dimensional real
Prior art date
Application number
PCT/CN2018/109728
Other languages
English (en)
French (fr)
Inventor
张�杰
于皓
李犇
张涧
张卓博
Original Assignee
阳光财产保险股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 阳光财产保险股份有限公司 filed Critical 阳光财产保险股份有限公司
Publication of WO2019137049A1 publication Critical patent/WO2019137049A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F21/60—Protecting data
    • G06F21/602—Providing cryptographic facilities or services
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F21/60—Protecting data
    • G06F21/62—Protecting access to data via a platform, e.g. using keys or access control rules
    • G06F21/6218—Protecting access to data via a platform, e.g. using keys or access control rules to a system of files or objects, e.g. local or distributed file system or database
    • G06F21/6245—Protecting personal data, e.g. for financial or medical purposes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/03—Credit; Loans; Processing thereof

Definitions

  • the present application relates to the technical field of the Internet, and in particular, to a method, device, electronic device, and computer storage medium based on information sharing.
  • Labeling The granularity of the original data is coarsened, and the level value is output, for example, the customer's monthly income is divided into real numbers from a real number. This practice affects the accuracy of data utilization and thus the prediction effect.
  • Encryption and decryption technology Encrypts the original data, which makes the data lose the sorting ability and the quantization ability after encryption, and thus cannot be applied to the prediction system.
  • the existing information sharing-based prediction method has the technical problems of poor data utilization accuracy, data loss sorting ability and quantization ability under the premise of protecting customer privacy data.
  • the purpose of the present application includes providing an information sharing-based prediction method, apparatus, electronic device, and computer storage medium to alleviate existing data sharing-based prediction methods in the presence of data to protect customer privacy data. Utilizing poor precision, the data loses the technical ability of sorting ability and quantization ability.
  • the embodiment of the present application provides a prediction method based on information sharing, which is applied to a client, and the method includes:
  • the server sends the multi-dimensional real number vector to the server, wherein after receiving the multi-dimensional real number vector sent by the plurality of clients, the server performs prediction according to the multi-dimensional real number vector and the updated weight corresponding to the multi-dimensional real number vector, and obtains a prediction result.
  • the embodiment of the present application provides a first possible implementation manner of the first aspect, wherein the client requirement includes: a data conversion frequency, a data conversion time, and a data conversion rule.
  • the embodiment of the present application provides a second possible implementation manner of the first aspect, wherein performing data conversion on the multi-dimensional original data according to the client requirement, and obtaining a multi-dimensional real number vector includes:
  • the embodiment of the present application provides a third possible implementation manner of the first aspect, wherein the sending the multi-dimensional real number vector to the server includes:
  • the multi-dimensional real number vector is sent to the server based on the connection relationship.
  • the embodiment of the present application provides a fourth possible implementation manner of the first aspect, wherein the joint modeling prediction model is capable of updating a weight corresponding to the multi-dimensional real number vector according to the data conversion rule. Obtaining an updated weight corresponding to the multidimensional real number vector.
  • the embodiment of the present application further provides a prediction method based on information sharing, which is applied to a server, and the method includes:
  • each client sends a multi-dimensional real number vector
  • the multi-dimensional real number vector is obtained by converting the multi-dimensional original data according to the client requirement by the corresponding client
  • the prediction is performed based on the updated weights corresponding to the multi-dimensional real number vector and the multi-dimensional real number vector, and the prediction result is obtained.
  • the embodiment of the present application provides a first possible implementation manner of the second aspect, wherein, after obtaining the prediction result, the method further includes:
  • the embodiment of the present application provides a second possible implementation manner of the second aspect, where the client requirements include: data conversion frequency, data conversion time, and data conversion rule.
  • the embodiment of the present application provides a third possible implementation manner of the second aspect, wherein, before acquiring the joint modeling prediction model, the method further includes:
  • the initial model is trained based on the training samples to obtain the joint modeling prediction model.
  • the embodiment of the present application further provides a prediction device based on information sharing, where the device is disposed on a client, and the device includes:
  • a data acquisition module configured to acquire multi-dimensional raw data of the user, where the user is a user who has historical data on the client;
  • a data conversion module configured to perform data conversion on the multi-dimensional raw data according to a client requirement to obtain a multi-dimensional real number vector, wherein, after the multi-dimensional original data is data-converted, the weight corresponding to the multi-dimensional real number vector is based on joint modeling
  • the control of the prediction model is updated to obtain updated weights, and the real number vector of each dimension corresponds to an updated weight, and the joint modeling prediction model is a pre-built prediction model;
  • a sending module configured to send the multi-dimensional real number vector to a server, where the server receives the multi-dimensional real number vector sent by the plurality of clients, and the updated weight corresponding to the multi-dimensional real number vector and the multi-dimensional real number vector Make predictions and get predictions.
  • the embodiment of the present application provides a first possible implementation manner of the third aspect, where the client requirement includes: a data conversion frequency, a data conversion time, and a data conversion rule.
  • the embodiment of the present application provides a second possible implementation manner of the third aspect, where the data conversion module includes:
  • An obtaining unit configured to acquire a data conversion rule in the client requirement, where the data conversion rule is customized by the client, and the data conversion rule does not affect a historical distribution rule of the original data of each dimension;
  • a data conversion unit configured to perform data conversion on the multi-dimensional raw data according to the data conversion rule to obtain the multi-dimensional real number vector, wherein the data conversion rule is confidential to other clients.
  • the embodiment of the present application provides a second possible implementation manner of the third aspect, where the sending module includes:
  • a sending unit configured to send the multi-dimensional real number vector to the server based on the connection relationship.
  • the embodiment of the present application further provides a prediction device based on information sharing, where the device is disposed on a server, and the device includes:
  • the receiving module is configured to receive a multi-dimensional real number vector sent by the multiple clients, where each client sends a multi-dimensional real number vector, and the multi-dimensional real number vector converts the multi-dimensional original data according to the client requirement by the corresponding client owned;
  • the weight update module is configured to update the weight corresponding to the multi-dimensional real number vector according to the control of the joint modeling prediction model to obtain an updated weight, wherein the real number vector of each dimension corresponds to an updated weight;
  • the prediction module is configured to perform prediction based on the multi-dimensional real number vector and the updated weight corresponding to the multi-dimensional real number vector to obtain a prediction result.
  • the embodiment of the present application provides the first possible implementation manner of the fourth aspect, wherein, after obtaining the prediction result, the apparatus further includes:
  • a monitoring module configured to monitor stability of the predicted result and monitor stability of the multi-dimensional real number vector.
  • an embodiment of the present application further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program
  • an electronic device including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program
  • the embodiment of the present application further provides a computer storage medium, where the computer program is stored, and the computer executes the computer program to perform the steps of the method according to any one of the foregoing first aspects.
  • the embodiments of the present application provide the following beneficial effects:
  • the embodiment of the present application provides a method, a device, an electronic device, and a computer storage medium based on information sharing, and the method is applied to a client, including: acquiring multi-dimensional raw data of a user.
  • the user is a user who has historical data on the client; the data of the multi-dimensional raw data is converted according to the client requirement, and a multi-dimensional real number vector is obtained, wherein, when the multi-dimensional original data is converted, the weight corresponding to the multi-dimensional real vector is based on the joint
  • the control of the modeling prediction model is updated to obtain updated weights, and the real number vector of each dimension corresponds to an updated weight, and the joint modeling prediction model is a pre-built prediction model;
  • the multi-dimensional real number vector is sent to the server, wherein After receiving the multi-dimensional real number vector sent by multiple clients, the server performs prediction based on the updated weights corresponding to the multi-dimensional real number vector and the multi-dimensional real number vector, and obtains the prediction result.
  • the existing information sharing-based prediction method generally thickens the granularity of the original data, and then directly outputs the gradation value, or encrypts the original data, and the encrypted data loses the sorting ability and the quantization ability, and thus cannot be used for subsequent Forecasting process.
  • the information sharing-based prediction method in the embodiment of the present application first acquires multi-dimensional raw data of the user, and then performs data conversion on the multi-dimensional original data according to the requirements of the client to obtain a multi-dimensional
  • the real number vector, and the joint modeling prediction model also controls the weight corresponding to the multidimensional real vector to update, and obtains the updated weight.
  • the updated weights corresponding to the multidimensional real vector and the multidimensional real vector sent by multiple clients. Forecast get prediction results.
  • the method fully applies the accuracy of the data, improves the prediction effect, and the client can perform data conversion on the original data to obtain a multi-dimensional real number vector, and avoids the attacker to the user under the premise of ensuring the data retaining ability and the quantization ability.
  • the reverse of the original data ensures the security of the user's private data, and alleviates the technical problems of the existing information sharing-based prediction method in the premise of protecting the customer's private data, such as poor data utilization accuracy, data loss sorting ability and quantization ability. .
  • FIG. 1 is a flowchart of a method for predicting information sharing according to an embodiment of the present application
  • FIG. 2 is a flowchart of performing multi-dimensional real number data conversion according to client requirements according to an embodiment of the present application
  • FIG. 3 is a flowchart of another method for predicting information sharing according to an embodiment of the present application.
  • FIG. 4 is a schematic diagram of a device for predicting information sharing based on an embodiment of the present application
  • FIG. 5 is a schematic diagram of another information sharing based prediction apparatus according to an embodiment of the present disclosure.
  • FIG. 6 is a schematic diagram of an information sharing-based prediction system according to an embodiment of the present application.
  • FIG. 7 is a schematic diagram of an electronic device according to an embodiment of the present application.
  • 11-data acquisition module 11-data acquisition module; 12-data conversion module; 13-transmission module; 21-receive module; 22-acquisition module; 23-weight update module; 24-prediction module.
  • Embodiment 1 is a diagrammatic representation of Embodiment 1:
  • a prediction method based on information sharing is applied to a client.
  • the method includes:
  • the client refers to a vendor client that has original data, and the client can be installed in a terminal device, where the manufacturer can be a P2P, a bank, a small loan company, a Taobao, a Baidu, an operator, etc.
  • the information sharing based prediction method provided by the present application may be performed by a processor of the terminal device where the client is located.
  • the original data includes various data, such as basic information of the user (including name, home address, work unit, etc.), user's income information, social security information, provident fund information, e-commerce shopping record information, call record information, etc.
  • the application examples do not specifically limit them.
  • the process of acquiring the user's multi-dimensional raw data is online acquisition of data.
  • the above historical data is a training sample used in constructing a joint modeling prediction model.
  • S104 Perform data conversion on the multi-dimensional raw data according to the client requirement, and obtain a multi-dimensional real number vector. wherein, when the multi-dimensional original data is converted, the weight corresponding to the multi-dimensional real number vector is updated according to the control of the joint modeling prediction model, and is updated. The weight of each dimension, the real number vector of each dimension corresponds to an updated weight, and the joint modeling prediction model is a pre-built prediction model;
  • the multi-dimensional raw data is converted according to the client's requirements, and a multi-dimensional real vector is obtained.
  • client requirements refer to the frequency of data conversion, data conversion time, data conversion rules, data conversion rules can be determined according to the needs of the client, and there is no specific limit. Each conversion can be for only one of the real vectors, or for multiple real vectors.
  • the weights corresponding to the multi-dimensional real numbers are updated according to the control of the joint modeling prediction model, and the updated weights corresponding to each real vector are obtained.
  • This data conversion is to prevent users from launching multi-dimensional raw data, resulting in the leakage of user privacy.
  • the joint modeling prediction model is a pre-built prediction model, which is a model built offline.
  • the user of the forecast result is a manufacturer, has its own product, wants to find its potential users by establishing a marketing model, or the user who predicts the result already has a user, and wants to predict the performance ability of the user by establishing a credit scoring model, except
  • the predictive model may also be an anti-fraud model, an application scoring model, a behavior scoring model, a collection model, and the like. If you want to build this model, the user will need to find multiple vendors to build the above model without own data or insufficient data.
  • the initial model is trained to obtain a joint modeling prediction model.
  • the model is a machine learning model trained with training samples.
  • the last vendor D has data from x9001 to x9001.
  • the xi in the model corresponds to the data source.
  • x5001 is the data provided by the manufacturer C, and the corresponding data is also corresponding.
  • the x5001 is the home appliance data when modeling, then the x5001 is also used when online. Home appliance data.
  • S106 Send the multi-dimensional real number vector to the server, where the server receives the multi-dimensional real number vector sent by the multiple clients, and performs prediction according to the updated weight corresponding to the multi-dimensional real number vector and the multi-dimensional real number vector to obtain a prediction result.
  • the client After the client obtains the multidimensional real vector, the client sends the multidimensional real vector to the server. After receiving the multi-dimensional real number vector sent by the multiple clients, the server receives the multi-dimensional real number vector sent by the four ABCD vendors on the line, and then performs the prediction according to the updated weight corresponding to the multi-dimensional real number vector and the multi-dimensional real number vector. , you can get the prediction results.
  • the existing information sharing-based prediction method generally thickens the granularity of the original data, and then directly outputs the gradation value, or encrypts the original data, and the encrypted data loses the sorting ability and the quantization ability, and thus cannot be used for subsequent Forecasting process.
  • the information sharing-based prediction method in the embodiment of the present application first acquires multi-dimensional raw data of the user, and then performs data conversion on the multi-dimensional original data according to the requirements of the client to obtain a multi-dimensional
  • the real number vector, and the joint modeling prediction model also controls the weight corresponding to the multidimensional real vector to update, and obtains the updated weight.
  • the updated weights corresponding to the multidimensional real vector and the multidimensional real vector sent by multiple clients. Forecast get prediction results.
  • the method fully applies the accuracy of the data, improves the prediction effect, and the client can perform data conversion on the original data to obtain a multi-dimensional real number vector, and avoids the attacker to the user under the premise of ensuring the data retaining ability and the quantization ability.
  • the reverse of the original data ensures the security of the user's private data, and alleviates the technical problems of the existing information sharing-based prediction method in the premise of protecting the customer's private data, such as poor data utilization accuracy, data loss sorting ability and quantization ability. .
  • the data transformation rules are customized by the client and may be inconsistent with the rules for establishing a joint modeling prediction model.
  • the multi-dimensional raw data provided by client A is the e-commerce shopping record (such as spending some money to buy a certain product in a certain period)
  • the multi-dimensional raw data is turned into a real vector, such as
  • the first dimension represents the consumption amount of the household appliance category
  • the second dimension represents the consumption amount of the clothing category
  • the third dimension represents the consumption amount of the cosmetics category, etc.
  • the amount can be converted into a real number between 0 and 1
  • the data conversion rule for example, the order of the two dimensions is changed, and the amount is converted into a real number between 0 and 100, etc., which is not specifically limited in the embodiment of the present application.
  • the data conversion rules do not affect the historical distribution of raw data for each dimension.
  • the monthly income generally shows a ⁇ distribution. After the data conversion rules are changed, the monthly income still follows the ⁇ distribution and does not affect the density distribution curve of each dimension.
  • Each change in the conversion rule is accompanied by an update of the weight of each dimension. That is, the joint modeling prediction model can update the weight corresponding to the multidimensional real number vector according to the data conversion rule, and obtain the updated weight corresponding to the multidimensional real number vector.
  • the data conversion rules of the A vendor are unknown to the B vendor.
  • sending the multidimensional real number vector to the server includes:
  • the multi-dimensional original data is first transformed and transformed on the client, so that the user's private data is protected; when the server uses the converted data for prediction, according to the client data.
  • the conversion rules synchronously update the weight of the model so as not to affect the accuracy and prediction effect of the data.
  • the client can initiate and change the rules of the multi-dimensional raw data at any time, which can further increase the possibility of the attacker to infer the private data.
  • the present application can encrypt the original data of the joint modeling participants to prevent reverse cracking while keeping the data usage efficiency from being lowered.
  • Embodiment 2 is a diagrammatic representation of Embodiment 1:
  • a prediction method based on information sharing is applied to a server. Referring to FIG. 3, the method includes:
  • S302. Receive a multi-dimensional real number vector sent by multiple clients, where each client sends a multi-dimensional real number vector, and the multi-dimensional real number vector is obtained by converting the multi-dimensional original data according to the client requirement by the corresponding client;
  • the method further includes:
  • the joint modeling prediction model is trained, and the raw data provided by the manufacturer is examined.
  • the client requirements include: data conversion frequency, data conversion time, and data conversion rules.
  • the joint modeling prediction model before acquiring the joint modeling prediction model, it also includes:
  • the initial model is trained based on the training samples to obtain a joint modeling prediction model.
  • Embodiment 3 is a diagrammatic representation of Embodiment 3
  • a prediction device based on information sharing the device is disposed at a client, and referring to FIG. 4, the device includes:
  • the data obtaining module 11 is configured to acquire multi-dimensional raw data of the user, where the user is a user who has historical data on the client;
  • the data conversion module 12 is configured to perform data conversion on the multi-dimensional raw data according to the client requirement, and obtain a multi-dimensional real number vector. wherein, when the multi-dimensional original data is converted, the weight corresponding to the multi-dimensional real number vector is controlled according to the joint modeling prediction model. Update, get the updated weight, the real number vector of each dimension corresponds to an updated weight, and the joint modeling prediction model is a pre-built prediction model;
  • the sending module 13 is configured to send the multi-dimensional real number vector to the server, wherein after receiving the multi-dimensional real number vector sent by the plurality of clients, the server performs prediction according to the updated weight corresponding to the multi-dimensional real number vector and the multi-dimensional real number vector to obtain a prediction result.
  • the multi-dimensional raw data of the user is first acquired, and then the multi-dimensional original data is converted according to the requirements of the client to obtain a multi-dimensional real number vector, and the joint modeling prediction model simultaneously controls the multi-dimensional
  • the weight corresponding to the real vector is updated to obtain the updated weight.
  • the updated weight is corresponding to the multi-dimensional real vector and the multi-dimensional real vector sent by the plurality of clients, and the prediction result is obtained.
  • the device fully applies the accuracy of the data, improves the prediction effect, and the client can perform data conversion on the original data to obtain a multi-dimensional real number vector, and avoids the attacker to the user under the premise of ensuring the data retaining ability and the quantization ability.
  • the reverse of the original data ensures the security of the user's private data, which alleviates the technical problems of the existing data sharing-based prediction device, which has poor data utilization accuracy, data loss sorting ability and quantization ability under the premise of protecting customer privacy data. .
  • the client requirements include: data conversion frequency, data conversion time, and data conversion rules.
  • the data conversion module comprises:
  • the obtaining unit is configured to obtain a data conversion rule in the client requirement, wherein the data conversion rule is customized by the client, and the data conversion rule does not affect the historical distribution rule of the original data of each dimension;
  • the data conversion unit is configured to perform data conversion on the multi-dimensional original data according to the data conversion rule to obtain a multi-dimensional real number vector, wherein the data conversion rule is confidential to other clients.
  • the sending module includes:
  • Establish a unit configured to establish a connection relationship with the server
  • a sending unit configured to send the multidimensional real number vector to the server based on the connection relationship.
  • the joint modeling prediction model can update the weight corresponding to the multi-dimensional real number vector according to the data conversion rule, and obtain the updated weight corresponding to the multi-dimensional real vector.
  • Embodiment 4 is a diagrammatic representation of Embodiment 4:
  • a prediction device based on information sharing the device is disposed on a server, and referring to FIG. 5, the device includes:
  • the receiving module 21 is configured to receive a multi-dimensional real number vector sent by multiple clients, where each client sends a multi-dimensional real number vector, and the multi-dimensional real number vector converts the multi-dimensional original data according to the client requirement by the corresponding client. of;
  • the obtaining module 22 is configured to acquire a joint modeling prediction model
  • the weight update module 23 is configured to update the weight corresponding to the multi-dimensional real number vector according to the control of the joint modeling prediction model to obtain an updated weight, wherein the real number vector of each dimension corresponds to an updated weight;
  • the prediction module 24 is configured to perform prediction based on the updated weights corresponding to the multi-dimensional real number vector and the multi-dimensional real number vector to obtain a prediction result.
  • the device further includes:
  • a monitoring module configured to monitor the stability of predicted results and to monitor the stability of multidimensional real numbers.
  • the client requirements include: data conversion frequency, data conversion time, and data conversion rules.
  • the device is further configured to:
  • the initial model is trained based on the training samples to obtain a joint modeling prediction model.
  • FIG. 6 is a schematic diagram of a prediction system based on information sharing.
  • the weight update module may be on the server side or on the client side.
  • the embodiment of the present application does not specifically limit the data transmission direction. .
  • Embodiment 5 is a diagrammatic representation of Embodiment 5:
  • the terminal device includes the client described in the foregoing Embodiment 1, and the terminal device includes: a processor 70, a memory 71, a bus 72, and a communication interface 77.
  • the processor 70, the communication interface 77 and the memory 71 are connected by a bus 72; the processor 70 is configured to execute an executable module, such as a computer program, stored in the memory 71.
  • the steps of the method as described in the first embodiment of the method are performed by the processor, and the method includes: acquiring the multi-dimensional raw data of the user, where the user is a user who has historical data on the client; The multi-dimensional raw data is subjected to data conversion to obtain a multi-dimensional real number vector, wherein, after the multi-dimensional raw data is subjected to data conversion, the weight corresponding to the multi-dimensional real number vector is updated according to the control of the joint modeling prediction model, and the updated Weight, the real number vector of each dimension corresponds to an updated weight, the joint modeling prediction model is a pre-built prediction model; the multi-dimensional real number vector is sent to a server, wherein the server receives multiple clients to send After the multi-dimensional real number vector, the prediction is performed according to the updated weights corresponding to the multi-dimensional real number vector and the multi-dimensional real number vector, and the prediction result is obtained.
  • the memory 71 may include a high speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory.
  • RAM high speed random access memory
  • non-volatile memory such as at least one disk memory.
  • the communication connection between the system network element and at least one other network element is implemented by at least one communication interface 77 (which may be wired or wireless), and may use an Internet, a wide area network, a local network, a metropolitan area network, or the like.
  • the bus 72 can be an ISA bus, a PCI bus, or an EISA bus.
  • the bus can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, only one double-headed arrow is shown in Figure 7, but it does not mean that there is only one bus or one type of bus.
  • the memory 71 is configured to store a program, and the processor 70 executes the program after receiving the execution instruction.
  • the method executed by the device defined by the flow process disclosed in any of the foregoing embodiments of the present application may be configured in the processor 70. Or implemented by the processor 70.
  • Processor 70 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method may be completed by an integrated logic circuit of hardware in the processor 70 or an instruction in the form of software.
  • the processor 70 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP processor, etc.), or a digital signal processor (DSP). ), Application Specific Integrated Circuit (ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware component.
  • the methods, steps, and logical block diagrams disclosed in the embodiments of the present application can be implemented or executed.
  • the general purpose processor may be a microprocessor or the processor or any conventional processor or the like.
  • the steps of the method disclosed in the embodiments of the present application may be directly implemented by the hardware decoding processor, or may be performed by a combination of hardware and software modules in the decoding processor.
  • the software module can be located in a conventional storage medium such as random access memory, flash memory, read only memory, programmable read only memory or electrically erasable programmable memory, registers, and the like.
  • the storage medium is located in the memory 71, and the processor 70 reads the information in the memory 71 and performs the steps of the above method in combination with its hardware.
  • a computer storage medium having stored thereon a computer program, the computer executing the method of any of the above method embodiments step.
  • the computer program product of the information sharing-based prediction method, device, electronic device and computer storage medium provided by the embodiment of the present application includes a computer readable storage medium storing program code, and the program code includes instructions configurable to execute
  • the terms “installation”, “connected”, and “connected” are to be understood broadly, and may be a fixed connection or a detachable connection, unless explicitly stated and defined otherwise. , or connected integrally; may be mechanical connection or electrical connection; may be directly connected, or may be indirectly connected through an intermediate medium, and may be internal communication between the two elements.
  • installation may be a fixed connection or a detachable connection, unless explicitly stated and defined otherwise.
  • connected integrally may be mechanical connection or electrical connection; may be directly connected, or may be indirectly connected through an intermediate medium, and may be internal communication between the two elements.
  • the functions may be stored in a computer readable storage medium if implemented in the form of a software functional unit and sold or used as a standalone product.
  • the technical solution of the present application which is essential or contributes to the prior art, or a part of the technical solution, may be embodied in the form of a software product, which is stored in a storage medium, including
  • the instructions are used to cause a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present application.
  • the foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, and the like. .
  • the information sharing-based prediction method, apparatus, electronic device and computer storage medium provided by the embodiments of the present application fully utilize the accuracy of data usage, improve the prediction effect, and the client can perform data conversion on the original data to obtain a multi-dimensional
  • the real number vector under the premise of ensuring the data retention ordering ability and the quantification ability, avoids the attacker's reverse push on the user's original data and ensures the security of the user's private data.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Bioethics (AREA)
  • General Health & Medical Sciences (AREA)
  • Strategic Management (AREA)
  • Health & Medical Sciences (AREA)
  • Economics (AREA)
  • Computer Security & Cryptography (AREA)
  • General Engineering & Computer Science (AREA)
  • Development Economics (AREA)
  • Software Systems (AREA)
  • General Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Computer Hardware Design (AREA)
  • Finance (AREA)
  • Marketing (AREA)
  • Accounting & Taxation (AREA)
  • Game Theory and Decision Science (AREA)
  • Tourism & Hospitality (AREA)
  • Quality & Reliability (AREA)
  • Medical Informatics (AREA)
  • Databases & Information Systems (AREA)
  • Operations Research (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Technology Law (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

本申请提供了一种基于信息共享的预测方法、装置、电子设备及计算机存储介质，该方法包括：获取多维原始数据；按照客户端需求对多维原始数据进行数据转换，得到多维实数向量，数据转换中，多维实数向量对应的权重随之更新，得到更新后的权重；服务器接收多个客户端发送的多维实数向量，根据多维实数向量和其对应的更新后的权重进行预测，得到预测结果。该方法充分应用了数据的使用精度，提高了预测效果，客户端对原始数据进行数据转换，得到多维实数向量，在保证数据保持排序能力和量化能力的前提下，避免了攻击者对用户原始数据的反推，保证了用户隐私数据的安全，缓解了现有的方法在保护客户隐私数据的前提下存在数据利用精度差，数据丧失排序能力和量化能力的技术问题。

Description

基于信息共享的预测方法、装置、电子设备及计算机存储介质
相关申请的交叉引用
本申请要求于2018年1月10日提交中国专利局的优先权号为2018100250734、名称为“基于信息共享的预测方法及装置”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及互联网的技术领域，尤其是涉及一种基于信息共享的预测方法、装置、电子设备及计算机存储介质。
背景技术
互联网金融近几年得到了飞速发展，各类金融科技公司呈现百家争鸣、百花齐放的局面，伴随着产业的欣荣发展，信息共享问题随之而来，目前没有哪家公司能够掌握风控所需的全部数据，因此经常会有多家公司联合建立风险预测模型的需求。与此同时，客户隐私数据受到法律保护，未经用户授权的情况下，客户隐私数据不得交换、共享。
现有技术中，对于多家公司联合建立风险预测模型(即基于信息共享的方法联合建立风险预测模型)存在的已有的处理方式主要有以下几种，然而这些技术均存在一定的不足：
标签化处理：将原始数据的粒度变粗，输出等级值，比如将客户的月收入从实数值划分为有限的几档。这种做法会影响数据利用的精度，从而影响预测效果。
加密解密技术：对原始数据进行加密，该方法使得数据在加密之后丧失了排序能力和量化能力，从而不能应用于预测系统。
综上，现有的基于信息共享的预测方法在保护客户隐私数据的前提下存在数据利用精度差，数据丧失排序能力和量化能力的技术问题。
申请内容
有鉴于此，本申请的目的包括，提供一种基于信息共享的预测方法、装置、电子设备及计算机存储介质，以缓解现有的基于信息共享的预测方法在保护客户隐私数据的前提下存在数据利用精度差，数据丧失排序能力和量化能力的技术问题。
第一方面，本申请实施例提供了一种基于信息共享的预测方法，应用于客户端，所述方法包括：
获取用户的多维原始数据，其中，所述用户为在客户端存在历史数据的用户；
按照客户端需求对所述多维原始数据进行数据转换，得到多维实数向量，其中，当所述多维原始数据进行数据转换后，所述多维实数向量对应的权重根据联合建模预测模 型的控制进行更新，得到更新后的权重，每一个维度的实数向量对应一个更新后的权重，所述联合建模预测模型为预先构建的预测模型；
将所述多维实数向量发送至服务器，其中，所述服务器接收多个客户端发送的多维实数向量后，根据所述多维实数向量和所述多维实数向量对应的更新后的权重进行预测，得到预测结果。
结合第一方面，本申请实施例提供了第一方面的第一种可能的实施方式，其中，所述客户端需求包括：数据转换频次，数据转换时间，数据转换规则。
结合第一方面，本申请实施例提供了第一方面的第二种可能的实施方式，其中，按照所述客户端需求对所述多维原始数据进行数据转换，得到多维实数向量包括：
获取所述客户端需求中的数据转换规则，其中，所述数据转换规则由所述客户端自定义，所述数据转换规则不影响每一维度原始数据的历史分布规律；
根据所述数据转换规则对所述多维原始数据进行数据转换，得到所述多维实数向量，其中，所述数据转换规则对于其它客户端保密。
结合第一方面，本申请实施例提供了第一方面的第三种可能的实施方式，其中，将所述多维实数向量发送至服务器包括：
建立与所述服务器的连接关系；
基于所述连接关系将所述多维实数向量发送至所述服务器。
结合第一方面，本申请实施例提供了第一方面的第四种可能的实施方式，其中，所述联合建模预测模型能够根据所述数据转换规则对所述多维实数向量对应的权重进行更新，得到所述多维实数向量对应的更新后的权重。
第二方面，本申请实施例还提供了一种基于信息共享的预测方法，应用于服务器，所述方法包括：
接收多个客户端发送的多维实数向量，其中，每个客户端发送一个多维实数向量，所述多维实数向量为其对应的客户端按照客户端需求对多维原始数据进行数据转换得到的；
获取联合建模预测模型；
根据所述联合建模预测模型的控制对所述多维实数向量对应的权重进行更新，得到更新后的权重，其中，每一个维度的实数向量对应一个更新后的权重；
基于所述多维实数向量和所述多维实数向量对应的更新后的权重进行预测，得到预测结果。
结合第二方面，本申请实施例提供了第二方面的第一种可能的实施方式，其中，在得到所述预测结果后，所述方法还包括：
监控所述预测结果的稳定性，以及监控所述多维实数向量的稳定性。
结合第二方面，本申请实施例提供了第二方面的第二种可能的实施方式，其中，所述客户端需求包括：数据转换频次，数据转换时间，数据转换规则。
结合第二方面，本申请实施例提供了第二方面的第三种可能的实施方式，其中，在获取联合建模预测模型之前还包括：
构建所述联合建模预测模型，包括：
获取初始模型和所述多个客户端提供的训练样本，其中，所述训练样本为所述多个客户端的历史数据；
基于所述训练样本对所述初始模型进行训练，得到所述联合建模预测模型。
第三方面，本申请实施例还提供了一种基于信息共享的预测装置，所述装置设置于客户端，所述装置包括：
数据获取模块，配置成获取用户的多维原始数据，其中，所述用户为在客户端存在历史数据的用户；
数据转换模块，配置成按照客户端需求对所述多维原始数据进行数据转换，得到多维实数向量，其中，当所述多维原始数据进行数据转换后，所述多维实数向量对应的权重根据联合建模预测模型的控制进行更新，得到更新后的权重，每一个维度的实数向量对应一个更新后的权重，所述联合建模预测模型为预先构建的预测模型；
发送模块，配置成将所述多维实数向量发送至服务器，其中，所述服务器接收多个客户端发送的多维实数向量后，根据所述多维实数向量和所述多维实数向量对应的更新后的权重进行预测，得到预测结果。
结合第三方面，本申请实施例提供了第三方面的第一种可能的实施方式，其中，所述客户端需求包括：数据转换频次，数据转换时间，数据转换规则。
结合第三方面，本申请实施例提供了第三方面的第二种可能的实施方式，其中，所述数据转换模块包括：
获取单元，配置成获取所述客户端需求中的数据转换规则，其中，所述数据转换规则由所述客户端自定义，所述数据转换规则不影响每一维度原始数据的历史分布规律；
数据转换单元，配置成根据所述数据转换规则对所述多维原始数据进行数据转换，得到所述多维实数向量，其中，所述数据转换规则对于其它客户端保密。
结合第三方面，本申请实施例提供了第三方面的第二种可能的实施方式，其中，所述发送模块包括：
建立单元，配置成建立与所述服务器的连接关系；
发送单元，配置成基于所述连接关系将所述多维实数向量发送至所述服务器。
第四方面，本申请实施例还提供了一种基于信息共享的预测装置，所述装置设置于服务器，所述装置包括：
接收模块，配置成接收多个客户端发送的多维实数向量，其中，每个客户端发送一个多维实数向量，所述多维实数向量为其对应的客户端按照客户端需求对多维原始数据进行数据转换得到的；
获取模块，配置成获取联合建模预测模型；
权重更新模块，配置成根据所述联合建模预测模型的控制对所述多维实数向量对应的权重进行更新，得到更新后的权重，其中，每一个维度的实数向量对应一个更新后的权重；
预测模块，配置成基于所述多维实数向量和所述多维实数向量对应的更新后的权重进行预测，得到预测结果。
结合第四方面，本申请实施例提供了第四方面的第一种可能的实施方式，其中，在得到所述预测结果后，所述装置还包括：
监控模块，配置成监控所述预测结果的稳定性，以及监控所述多维实数向量的稳定性。
第五方面，本申请实施例还提供了一种电子设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述处理器执行所述计算机程序时实现上述第一方面中任一项所述的方法。
第六方面，本申请实施例还提供了一种计算机存储介质，其上存储有计算机程序，所述计算机运行所述计算机程序时执行上述第一方面中任一项所述的方法的步骤。
本申请实施例带来了以下有益效果：本申请实施例提供了一种基于信息共享的预测方法、装置、电子设备及计算机存储介质，该方法应用于客户端，包括：获取用户的多维原始数据，其中，用户为在客户端存在历史数据的用户；按照客户端需求对多维原始数据进行数据转换，得到多维实数向量，其中，当多维原始数据进行数据转换后，多维实数向量对应的权重根据联合建模预测模型的控制进行更新，得到更新后的权重，每一个维度的实数向量对应一个更新后的权重，联合建模预测模型为预先构建的预测模型；将多维实数向量发送至服务器，其中，服务器接收多个客户端发送的多维实数向量后，根据多维实数向量和多维实数向量对应的更新后的权重进行预测，得到预测结果。
现有的基于信息共享的预测方法一般是将原始数据的粒度变粗，然后直接输出等级值，或者将原始数据进行加密，加密后的数据丧失了排序能力和量化能力，进而无法用于后续的预测过程。与现有的基于信息共享的预测方法相比，本申请实施例中的基于信息共享的预测方法中，先获取用户的多维原始数据，然后根据客户端的需求对多维原始 数据进行数据转换，得到多维实数向量，并且联合建模预测模型同时会控制多维实数向量对应的权重进行更新，得到更新后的权重，最终，根据多个客户端发送的多维实数向量和多维实数向量对应的更新后的权重进行预测，得到预测结果。该方法充分应用了数据的使用精度，提高了预测效果，并且客户端能够对原始数据进行数据转换，得到多维实数向量，在保证数据保持排序能力和量化能力的前提下，避免了攻击者对用户原始数据的反推，保证了用户隐私数据的安全性，缓解了现有的基于信息共享的预测方法在保护客户隐私数据的前提下存在数据利用精度差，数据丧失排序能力和量化能力的技术问题。
本申请的其他特征和优点将在随后的说明书中阐述，并且，部分地从说明书中变得显而易见，或者通过实施本申请而了解。本申请的目的和其他优点在说明书、权利要求书以及附图中所特别指出的结构来实现和获得。
为使本申请的上述目的、特征和优点能更明显易懂，下文特举较佳实施例，并配合所附附图，作详细说明如下。
附图说明
为了更清楚地说明本申请具体实施方式或现有技术中的技术方案，下面将对具体实施方式或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图是本申请的一些实施方式，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1为本申请实施例提供的一种基于信息共享的预测方法的流程图；
图2为本申请实施例提供的按照客户端需求对多维原始数据进行数据转换，得到多维实数向量的流程图；
图3为本申请实施例提供的另一种基于信息共享的预测方法的流程图；
图4为本申请实施例提供的一种基于信息共享的预测装置的示意图；
图5为本申请实施例提供的另一种基于信息共享的预测装置的示意图；
图6为本申请实施例提供的基于信息共享的预测系统的示意图；
图7为本申请实施例提供的一种电子设备的示意图。
图标：
11-数据获取模块；12-数据转换模块；13-发送模块；21-接收模块；22-获取模块；23-权重更新模块；24-预测模块。
具体实施方式
为使本申请实施例的目的、技术方案和优点更加清楚，下面将结合附图对本申请的技术方案进行清楚、完整地描述，显然，所描述的实施例是本申请一部分实施例，而不是全 部的实施例。基于本申请中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
为便于对本实施例进行理解，首先对本申请实施例所公开的一种基于信息共享的预测方法进行详细介绍。
实施例一：
一种基于信息共享的预测方法，应用于客户端，参考图1，该方法包括：
S102、获取用户的多维原始数据，其中，用户为在客户端存在历史数据的用户；
在本申请实施例中，客户端是指拥有原始数据的厂商客户端，该客户端可以安装在终端设备，其中，厂商可以为P2P，银行，小贷公司，淘宝，百度，运营商等，本申请实施例对其不做具体限制。在本实施例中，可以通过客户端所在终端设备的处理器执行本申请所提供的基于信息共享的预测方法。
原始数据包括各种各样的数据，比如用户的基本信息(包括姓名，家庭住址，工作单位等)，用户的收入信息，社保信息，公积金信息，电商购物记录信息，通话记录信息等，本申请实施例对其不做具体限制。
该获取用户的多维原始数据为线上获取数据的过程。
上述历史数据为构建联合建模预测模型时用到的训练样本。
S104、按照客户端需求对多维原始数据进行数据转换，得到多维实数向量，其中，当多维原始数据进行数据转换后，多维实数向量对应的权重根据联合建模预测模型的控制进行更新，得到更新后的权重，每一个维度的实数向量对应一个更新后的权重，联合建模预测模型为预先构建的预测模型；
在得到多维原始数据后，按照客户端需求对多维原始数据进行转换，得到多维实数向量。所谓的客户端需求是指数据转换频次，数据转换时间，数据转换规则，数据转换规则可以根据客户端的需要而定，并没有具体限制。每次转换可以只针对其中的一个实数向量，也可以针对多个实数向量。
当对多维原始数据进行数据转换后，多维实数向量对应的权重根据联合建模预测模型的控制进行更新，得到每个实数向量对应的更新后的权重。
该数据转换的目的是为了防止使用者反推出多维原始数据，造成对用户隐私的泄密。
联合建模预测模型为预先构建的预测模型，该模型为线下建立的模型。比如预测结果的使用方是一家厂商，有自己的产品，想通过建立营销模型找到其潜在用户，或者预测结果的使用方已经有了用户，想通过建立信用评分模型预测该用户的履约能力，除了上述营 销模型和信用评分模型外，预测模型还可能是反欺诈模型、申请评分模型、行为评分模型、催收模型等。如果要建立这个模型，使用方在没有自有数据或者自有数据不够的情况下，就需要找多家厂商，建立上述的模型。
构建联合建模预测模型，包括：
(1)获取初始模型和多个客户端提供的训练样本，其中，训练样本为多个客户端的历史数据；
(2)基于训练样本对初始模型进行训练，得到联合建模预测模型。
建模的过程中可以通过统计学习方法，事先选取一定数量的训练样本建立模型，也可以在没有训练样本的情况下依据专家经验设计这样一套模型。该预测模型的预测结果可能是关于实数向量与其对应的权重的函数(比如：y＝f(w1*x1+w2*x2+...+wn*xn)，其中，x1，x2，...，xn为n维实数向量的值，w1，w2，...，wn为n维实数向量对应的n个权重，也可能是概率分布函数，或是矩阵映射，或是神经网络结构等形式。
建模前需要对建模时用到的厂商提供的数据做一些数据调研，看一下数据的质量怎么样。比如说模型是用训练样本训练出来的机器学习模型，那样，可能需要自己先准备十万条样本，也就是十万个用户的数据，拿用户的唯一标识(比如身份证号，手机号码)，先要去其他厂商，让那些厂商为预测结果的使用方按照这些用户去准备他们拥有的数据。比如说有一家厂商A拥有的数据是x1到x200，另外一家厂商B拥有的数据是从x201到x5000，还有一家厂商C拥有的数据是x5001到x9000，最后一家厂商D拥有的数据是x9001到x100000，得到这些数据后，用这些数据去建立一个模型，通过统计学习方法，知道这十万条样本上面每个x相应的w是多少，就是机器学习的模型，通过历史的样本，让厂商ABCD把那些x给预测结果的使用方，然后，进行拟合，使得历史的真实的y与拟合出来的y之间的差尽可能的小，这样就能训练得到一个模型。
当训练得到一个模型后，将该模型放在服务器端，以进行线上的预测。
线下建立模型的方式千差万别，本申请实施例对其不做具体限制。
如果在线下建立模型的时候使用的为ABCD四家厂商的数据，建立得到模型后，把这个模型放到服务器上，在线上使用时，也只能使用ABCD四家厂商的数据。并且，模型中的xi对应好了数据来源，比如x5001就是厂商C提供的数据，具体也对应好了是哪类数据，比如建模时x5001为家电类数据，那么，线上使用时x5001也为家电类数据。
S106、将多维实数向量发送至服务器，其中，服务器接收多个客户端发送的多维实数向量后，根据多维实数向量和多维实数向量对应的更新后的权重进行预测，得到预测结果。
客户端在得到多维实数向量后，将多维实数向量发送至服务器。服务器接收多个客户 端发送的多维实数向量后，如步骤S104所述，服务器在线上接收ABCD四家厂商发送的多维实数向量后，根据多维实数向量和多维实数向量对应的更新后的权重进行预测，就能够得到预测结果。
现有的基于信息共享的预测方法一般是将原始数据的粒度变粗，然后直接输出等级值，或者将原始数据进行加密，加密后的数据丧失了排序能力和量化能力，进而无法用于后续的预测过程。与现有的基于信息共享的预测方法相比，本申请实施例中的基于信息共享的预测方法中，先获取用户的多维原始数据，然后根据客户端的需求对多维原始数据进行数据转换，得到多维实数向量，并且联合建模预测模型同时会控制多维实数向量对应的权重进行更新，得到更新后的权重，最终，根据多个客户端发送的多维实数向量和多维实数向量对应的更新后的权重进行预测，得到预测结果。该方法充分应用了数据的使用精度，提高了预测效果，并且客户端能够对原始数据进行数据转换，得到多维实数向量，在保证数据保持排序能力和量化能力的前提下，避免了攻击者对用户原始数据的反推，保证了用户隐私数据的安全性，缓解了现有的基于信息共享的预测方法在保护客户隐私数据的前提下存在数据利用精度差，数据丧失排序能力和量化能力的技术问题。
上述过程对基于信息共享的预测方法进行了简要描述，下面对其中涉及到的具体内容进行详细描述。
可选地，参考图2，按照客户端需求对多维原始数据进行数据转换，得到多维实数向量包括：
S201、获取客户端需求中的数据转换规则，其中，数据转换规则由客户端自定义，数据转换规则不影响每一维度原始数据的历史分布规律；
在一种实施方式中，数据转换规则由客户端自定义，可以与建立联合建模预测模型时的规则不一致。
比如，客户端A提供的多维原始数据为电商购物记录(如在某时期花了某些钱买了某个产品)，得到数据转换规则后，把多维的原始数据变成实数向量后，如第一维度代表家电类的消费金额，第二维度代表衣帽类的消费金额，第三维度代表化妆品类的消费金额等等，这些金额可以转换为0到1之间的实数，客户端可以改变数据转换规则，比如调换两个维度之间的顺序，将金额转换为0到100之间的实数等，本申请实施例对其不做具体限制。
另外，数据转换规则不影响每一维度原始数据的历史分布规律。比如，月收入总体上呈现出λ分布，数据转换规则改变后，月收入还是遵从λ分布，不影响每一维度的密度分布曲线。
每次转换规则的变化均伴随着每一维度的权重的更新。即联合建模预测模型能够根据数据转换规则对多维实数向量对应的权重进行更新，得到多维实数向量对应的更新后的权重。
S202、根据数据转换规则对多维原始数据进行数据转换，得到多维实数向量，其中，数据转换规则对于其它客户端保密。
比如，A厂商的数据转换规则对于B厂商说是未知的。
可选地，将多维实数向量发送至服务器包括：
(1)建立与服务器的连接关系；
(2)基于连接关系将多维实数向量发送至服务器。
本申请提供的基于信息共享的预测方法中，先在客户端对多维原始数据做变形转换，使得用户的隐私数据得到了保护；在服务器端使用转换后的数据做预测的时候，根据客户端的数据转换规则，同步的更新了模型的权重，从而不影响数据的使用精度和预测效果。另外，客户端对多维原始数据的数据转换可以随时发起、随时更改规则，这样可以更进一步加大了攻击者反向推测出隐私数据的可能性。
因此，本申请可以对联合建模参与方的原始数据做加密，防止逆向破解，同时保持数据的使用效率不降低。
实施例二：
一种基于信息共享的预测方法，应用于服务器，参考图3，该方法包括：
S302、接收多个客户端发送的多维实数向量，其中，每个客户端发送一个多维实数向量，多维实数向量为其对应的客户端按照客户端需求对多维原始数据进行数据转换得到的；
S304、获取联合建模预测模型；
S306、根据联合建模预测模型的控制对多维实数向量对应的权重进行更新，得到更新后的权重，其中，每一个维度的实数向量对应一个更新后的权重；
S308、基于多维实数向量和多维实数向量对应的更新后的权重进行预测，得到预测结果。
该实施例二中的具体内容可以参考实施例一中的描述，在此不再赘述。
可选地，在得到预测结果后，该方法还包括：
监控预测结果的稳定性，以及监控多维实数向量的稳定性。
另外，在得到预测结果后，还可以监控预测结果以及多维实数向量的稳定性，根据监控结果采取对应的措施。比如对联合建模预测模型进行训练，对厂商提供的原始数据进行 考察等。
可选地，客户端需求包括：数据转换频次，数据转换时间，数据转换规则。
可选地，在获取联合建模预测模型之前还包括：
构建联合建模预测模型，包括：
获取初始模型和多个客户端提供的训练样本，其中，训练样本为多个客户端的历史数据；
基于训练样本对初始模型进行训练，得到联合建模预测模型。
实施例三：
一种基于信息共享的预测装置，该装置设置于客户端，参考图4，该装置包括：
数据获取模块11，配置成获取用户的多维原始数据，其中，用户为在客户端存在历史数据的用户；
数据转换模块12，配置成按照客户端需求对多维原始数据进行数据转换，得到多维实数向量，其中，当多维原始数据进行数据转换后，多维实数向量对应的权重根据联合建模预测模型的控制进行更新，得到更新后的权重，每一个维度的实数向量对应一个更新后的权重，联合建模预测模型为预先构建的预测模型；
发送模块13，配置成将多维实数向量发送至服务器，其中，服务器接收多个客户端发送的多维实数向量后，根据多维实数向量和多维实数向量对应的更新后的权重进行预测，得到预测结果。
本申请实施例中的基于信息共享的预测装置中，先获取用户的多维原始数据，然后根据客户端的需求对多维原始数据进行数据转换，得到多维实数向量，并且联合建模预测模型同时会控制多维实数向量对应的权重进行更新，得到更新后的权重，最终，根据多个客户端发送的多维实数向量和多维实数向量对应的更新后的权重进行预测，得到预测结果。该装置充分应用了数据的使用精度，提高了预测效果，并且客户端能够对原始数据进行数据转换，得到多维实数向量，在保证数据保持排序能力和量化能力的前提下，避免了攻击者对用户原始数据的反推，保证了用户隐私数据的安全性，缓解了现有的基于信息共享的预测装置在保护客户隐私数据的前提下存在数据利用精度差，数据丧失排序能力和量化能力的技术问题。
可选地，客户端需求包括：数据转换频次，数据转换时间，数据转换规则。
可选地，数据转换模块包括：
获取单元，配置成获取客户端需求中的数据转换规则，其中，数据转换规则由客户端 自定义，数据转换规则不影响每一维度原始数据的历史分布规律；
数据转换单元，配置成根据数据转换规则对多维原始数据进行数据转换，得到多维实数向量，其中，数据转换规则对于其它客户端保密。
可选地，发送模块包括：
建立单元，配置成建立与服务器的连接关系；
发送单元，配置成基于连接关系将多维实数向量发送至服务器。
可选地，联合建模预测模型能够根据数据转换规则对多维实数向量对应的权重进行更新，得到多维实数向量对应的更新后的权重。
该实施例三中的具体内容可以参考实施例一中的具体描述，在此不再赘述。
实施例四：
一种基于信息共享的预测装置，该装置设置于服务器，参考图5，该装置包括：
接收模块21，配置成接收多个客户端发送的多维实数向量，其中，每个客户端发送一个多维实数向量，多维实数向量为其对应的客户端按照客户端需求对多维原始数据进行数据转换得到的；
获取模块22，配置成获取联合建模预测模型；
权重更新模块23，配置成根据联合建模预测模型的控制对多维实数向量对应的权重进行更新，得到更新后的权重，其中，每一个维度的实数向量对应一个更新后的权重；
预测模块24，配置成基于多维实数向量和多维实数向量对应的更新后的权重进行预测，得到预测结果。
该实施例四中的具体描述也可参考实施例一中的描述，在此不再赘述。
可选地，在得到预测结果后，该装置还包括：
监控模块，配置成监控预测结果的稳定性，以及监控多维实数向量的稳定性。
可选地，客户端需求包括：数据转换频次，数据转换时间，数据转换规则。
可选地，该装置还用于：
构建联合建模预测模型，包括：
获取初始模型和多个客户端提供的训练样本，其中，训练样本为多个客户端的历史数据；
基于训练样本对初始模型进行训练，得到联合建模预测模型。
图6给出了基于信息共享的预测系统的示意图，其中，权重更新模块可以在服务器端，也可以在客户端，本申请实施例对其不做具体限制，图中示出了数据的传输方向。
实施例五：
本申请实施例提供了一种终端设备，参考图7，该终端设备包括上述实施例一中所描述的客户端，该终端设备包括：处理器70，存储器71，总线72和通信接口77，处理器70、通信接口77和存储器71通过总线72连接；处理器70配置成执行存储器71中存储的可执行模块，例如计算机程序。处理器执行极端及程序时实现如方法实施例一中描述的方法的步骤，具体包括：获取用户的多维原始数据，其中，所述用户为在客户端存在历史数据的用户；按照客户端需求对所述多维原始数据进行数据转换，得到多维实数向量，其中，当所述多维原始数据进行数据转换后，所述多维实数向量对应的权重根据联合建模预测模型的控制进行更新，得到更新后的权重，每一个维度的实数向量对应一个更新后的权重，所述联合建模预测模型为预先构建的预测模型；将所述多维实数向量发送至服务器，其中，所述服务器接收多个客户端发送的多维实数向量后，根据所述多维实数向量和所述多维实数向量对应的更新后的权重进行预测，得到预测结果。
其中，存储器71可能包含高速随机存取存储器(RAM，Random Access Memory)，也可能还包括非不稳定的存储器(non-volatile memory)，例如至少一个磁盘存储器。通过至少一个通信接口77(可以是有线或者无线)实现该系统网元与至少一个其他网元之间的通信连接，可以使用互联网，广域网，本地网，城域网等。
总线72可以是ISA总线、PCI总线或EISA总线等。总线可以分为地址总线、数据总线、控制总线等。为便于表示，图7中仅用一个双向箭头表示，但并不表示仅有一根总线或一种类型的总线。
其中，存储器71配置成存储程序，处理器70在接收到执行指令后，执行程序，前述本申请实施例任一实施例揭示的流过程定义的装置所执行的方法可以应配置成处理器70中，或者由处理器70实现。
处理器70可能是一种集成电路芯片，具有信号的处理能力。在实现过程中，上述方法的各步骤可以通过处理器70中的硬件的集成逻辑电路或者软件形式的指令完成。上述的处理器70可以是通用处理器，包括中央处理器(Central Processing Unit，简称CPU)、网络处理器(Network Processor，简称NP)等；还可以是数字信号处理器(Digital Signal Processing，简称DSP)、专用集成电路(Application Specific Integrated Circuit，简称ASIC)、现成可编程门阵列(Field-Programmable Gate Array，简称FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。可以实现或者执行本申请实施例中的公开的各方法、步骤及逻辑框图。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。结合本申请实施例所公开的方法的步骤可以直接体现为硬件译码处理器执行完成，或者用 译码处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器71，处理器70读取存储器71中的信息，结合其硬件完成上述方法的步骤。
在本申请的另一个实施例中，还提供了一种计算机存储介质，其上存储有计算机程序，所述计算机运行所述计算机程序时执行上述方法实施例一中任一项所述的方法的步骤。
本申请实施例所提供的基于信息共享的预测方法、装置、电子设备及计算机存储介质的计算机程序产品，包括存储了程序代码的计算机可读存储介质，所述程序代码包括的指令可配置成执行前面方法实施例中所述的方法，具体实现可参见方法实施例，在此不再赘述。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统和装置的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
另外，在本申请实施例的描述中，除非另有明确的规定和限定，术语“安装”、“相连”、“连接”应做广义理解，例如，可以是固定连接，也可以是可拆卸连接，或一体地连接；可以是机械连接，也可以是电连接；可以是直接相连，也可以通过中间媒介间接相连，可以是两个元件内部的连通。对于本领域的普通技术人员而言，可以具体情况理解上述术语在本申请中的具体含义。
所述功能如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
在本申请的描述中，需要说明的是，术语“中心”、“上”、“下”、“左”、“右”、“竖直”、“水平”、“内”、“外”等指示的方位或位置关系为基于附图所示的方位或位置关系，仅是为了便于描述本申请和简化描述，而不是指示或暗示所指的装置或元件必须具有特定的方位、以特定的方位构造和操作，因此不能理解为对本申请的限制。此外，术语“第一”、“第二”、“第三”仅用于描述目的，而不能理解为指示或暗示相对重要性。
最后应说明的是：以上所述实施例，仅为本申请的具体实施方式，用以说明本申请的 技术方案，而非对其限制，本申请的保护范围并不局限于此，尽管参照前述实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，其依然可以对前述实施例所记载的技术方案进行修改或可轻易想到变化，或者对其中部分技术特征进行等同替换；而这些修改、变化或者替换，并不使相应技术方案的本质脱离本申请实施例技术方案的精神和范围，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应所述以权利要求的保护范围为准。
工业实用性
本申请实施例提供的基于信息共享的预测方法、装置、电子设备及计算机存储介质，该方法充分应用了数据的使用精度，提高了预测效果，并且客户端能够对原始数据进行数据转换，得到多维实数向量，在保证数据保持排序能力和量化能力的前提下，避免了攻击者对用户原始数据的反推，保证了用户隐私数据的安全性。

Claims (16)

  1. 一种基于信息共享的预测方法，其特征在于，应用于客户端，所述方法包括：
    获取用户的多维原始数据，其中，所述用户为在客户端存在历史数据的用户；
    按照客户端需求对所述多维原始数据进行数据转换，得到多维实数向量，其中，当所述多维原始数据进行数据转换后，所述多维实数向量对应的权重根据联合建模预测模型的控制进行更新，得到更新后的权重，每一个维度的实数向量对应一个更新后的权重，所述联合建模预测模型为预先构建的预测模型；
    将所述多维实数向量发送至服务器，其中，所述服务器接收多个客户端发送的多维实数向量后，根据所述多维实数向量和所述多维实数向量对应的更新后的权重进行预测，得到预测结果。
  2. 根据权利要求1所述的方法，其特征在于，所述客户端需求包括：数据转换频次，数据转换时间，数据转换规则。
  3. 根据权利要求1或2所述的方法，其特征在于，按照所述客户端需求对所述多维原始数据进行数据转换，得到多维实数向量包括：
    获取所述客户端需求中的数据转换规则，其中，所述数据转换规则由所述客户端自定义，所述数据转换规则不影响每一维度原始数据的历史分布规律；
    根据所述数据转换规则对所述多维原始数据进行数据转换，得到所述多维实数向量，其中，所述数据转换规则对于其它客户端保密。
  4. 根据权利要求1至3中任一项所述的方法，其特征在于，将所述多维实数向量发送至服务器包括：
    建立与所述服务器的连接关系；
    基于所述连接关系将所述多维实数向量发送至所述服务器。
  5. 根据权利要求2所述的方法，其特征在于，所述联合建模预测模型能够根据所述数据转换规则对所述多维实数向量对应的权重进行更新，得到所述多维实数向量对应的更新后的权重。
  6. 一种基于信息共享的预测方法，其特征在于，应用于服务器，所述方法包括：
    接收多个客户端发送的多维实数向量，其中，每个客户端发送一个多维实数向量，所述多维实数向量为其对应的客户端按照客户端需求对多维原始数据进行数据转换得到的；
    获取联合建模预测模型；
    根据所述联合建模预测模型的控制对所述多维实数向量对应的权重进行更新，得到更新后的权重，其中，每一个维度的实数向量对应一个更新后的权重；
    基于所述多维实数向量和所述多维实数向量对应的更新后的权重进行预测，得到预测结果。
  7. 根据权利要求6所述的方法，其特征在于，在得到所述预测结果后，所述方法还包括：
    监控所述预测结果的稳定性，以及监控所述多维实数向量的稳定性。
  8. 根据权利要求6所述的方法，其特征在于，所述客户端需求包括：数据转换频次，数据转换时间，数据转换规则。
  9. 根据权利要求6至8中任一项所述的方法，其特征在于，在获取联合建模预测模型之前还包括：
    构建所述联合建模预测模型，包括：
    获取初始模型和所述多个客户端提供的训练样本，其中，所述训练样本为所述多个客户端的历史数据；
    基于所述训练样本对所述初始模型进行训练，得到所述联合建模预测模型。
  10. 一种基于信息共享的预测装置，其特征在于，所述装置设置于客户端，所述装置包括：
    数据获取模块，配置成获取用户的多维原始数据，其中，所述用户为在客户端存在历史数据的用户；
    数据转换模块，配置成按照客户端需求对所述多维原始数据进行数据转换，得到多维实数向量，其中，当所述多维原始数据进行数据转换后，所述多维实数向量对应的权重根据联合建模预测模型的控制进行更新，得到更新后的权重，每一个维度的实数向量对应一个更新后的权重，所述联合建模预测模型为预先构建的预测模型；
    发送模块，配置成将所述多维实数向量发送至服务器，其中，所述服务器接收多个客户端发送的多维实数向量后，根据所述多维实数向量和所述多维实数向量对应的更新后的权重进行预测，得到预测结果。
  11. 根据权利要求10所述的装置，其特征在于，所述数据转换模块包括：
    获取单元，配置成获取所述客户端需求中的数据转换规则，其中，所述数据转换规则由所述客户端自定义，所述数据转换规则不影响每一维度原始数据的历史分布规律；
    数据转换单元，配置成根据所述数据转换规则对所述多维原始数据进行数据转换，得到所述多维实数向量，其中，所述数据转换规则对于其它客户端保密。
  12. 根据权利要求10或11所述的装置，其特征在于，所述发送模块包括：
    建立单元，配置成建立与所述服务器的连接关系；
    发送单元，配置成基于所述连接关系将所述多维实数向量发送至所述服务器。
  13. 一种基于信息共享的预测装置，其特征在于，所述装置设置于服务器，所述装置包括：
    接收模块，配置成接收多个客户端发送的多维实数向量，其中，每个客户端发送一个多维实数向量，所述多维实数向量为其对应的客户端按照客户端需求对多维原始数据进行数据转换得到的；
    获取模块，配置成获取联合建模预测模型；
    权重更新模块，配置成根据所述联合建模预测模型的控制对所述多维实数向量对应的权重进行更新，得到更新后的权重，其中，每一个维度的实数向量对应一个更新后的权重；
    预测模块，配置成基于所述多维实数向量和所述多维实数向量对应的更新后的权重进行预测，得到预测结果。
  14. 根据权利要求13所述的装置，其特征在于，在得到所述预测结果后，所述装置还包括：
    监控模块，配置成监控所述预测结果的稳定性，以及监控所述多维实数向量的稳定性。
  15. 一种电子设备，其特征在于，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述处理器执行所述计算机程序时实现上述权利要求1至9中任一项所述的方法。
  16. 一种计算机存储介质，其特征在于，其上存储有计算机程序，所述计算机运行所述计算机程序时执行上述权利要求1至9中任一项所述的方法的步骤。
PCT/CN2018/109728 2018-01-10 2018-10-10 基于信息共享的预测方法、装置、电子设备及计算机存储介质 WO2019137049A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810025073.4A CN108133294B (zh) 2018-01-10 2018-01-10 基于信息共享的预测方法及装置
CN201810025073.4 2018-01-10

Publications (1)

Publication Number Publication Date
WO2019137049A1 true WO2019137049A1 (zh) 2019-07-18

Family

ID=62400596

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/109728 WO2019137049A1 (zh) 2018-01-10 2018-10-10 基于信息共享的预测方法、装置、电子设备及计算机存储介质

Country Status (2)

Country Link
CN (1) CN108133294B (zh)
WO (1) WO2019137049A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111915368A (zh) * 2020-07-30 2020-11-10 上海数策软件股份有限公司 汽车行业客户id识别系统、方法及介质
CN113554476A (zh) * 2020-04-23 2021-10-26 京东数字科技控股有限公司 信用度预测模型的训练方法、系统、电子设备及存储介质

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108133294B (zh) * 2018-01-10 2020-12-04 阳光财产保险股份有限公司 基于信息共享的预测方法及装置
CN109635422B (zh) * 2018-12-07 2023-08-25 深圳前海微众银行股份有限公司 联合建模方法、装置、设备以及计算机可读存储介质
CN110969261B (zh) * 2019-11-29 2023-11-28 中国银行股份有限公司 基于加密算法的模型构建方法及相关设备
CN111367960A (zh) * 2020-02-25 2020-07-03 北京明略软件系统有限公司 一种实现数据处理的方法、装置、计算机存储介质及终端
CN111160573B (zh) * 2020-04-01 2020-06-30 支付宝(杭州)信息技术有限公司 保护数据隐私的双方联合训练业务预测模型的方法和装置

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103886203A (zh) * 2014-03-24 2014-06-25 美商天睿信息系统（北京）有限公司 一种基于指标预测的自动建模系统及其方法
CN104700155A (zh) * 2014-12-24 2015-06-10 天津南大通用数据技术股份有限公司 商业智能中的业务模型利用pmml实现预测的方法及系统
CN104794406A (zh) * 2015-03-18 2015-07-22 云南电网有限责任公司电力科学研究院 一种基于数据迷彩模型的私密数据保护方法
CN105894336A (zh) * 2016-05-25 2016-08-24 北京比邻弘科科技有限公司 一种基于移动互联网的大数据挖掘方法及系统
US20170061311A1 (en) * 2015-08-27 2017-03-02 Li Liu Method for providing data analysis service by a service provider to data owner and related data transformation method for preserving business confidential information of the data owner
CN107480549A (zh) * 2017-06-28 2017-12-15 银江股份有限公司 一种面向数据共享的敏感信息脱敏方法及系统
CN108133294A (zh) * 2018-01-10 2018-06-08 阳光财产保险股份有限公司 基于信息共享的预测方法及装置

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101079063A (zh) * 2007-06-25 2007-11-28 腾讯科技（深圳）有限公司 一种基于场景信息推送广告的方法、系统及设备
CN103198418A (zh) * 2013-03-15 2013-07-10 北京亿赞普网络技术有限公司 一种应用推荐方法和系统
CN106603293A (zh) * 2016-12-20 2017-04-26 南京邮电大学 虚拟网络环境下一种基于深度学习的网络故障诊断方法

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103886203A (zh) * 2014-03-24 2014-06-25 美商天睿信息系统（北京）有限公司 一种基于指标预测的自动建模系统及其方法
CN104700155A (zh) * 2014-12-24 2015-06-10 天津南大通用数据技术股份有限公司 商业智能中的业务模型利用pmml实现预测的方法及系统
CN104794406A (zh) * 2015-03-18 2015-07-22 云南电网有限责任公司电力科学研究院 一种基于数据迷彩模型的私密数据保护方法
US20170061311A1 (en) * 2015-08-27 2017-03-02 Li Liu Method for providing data analysis service by a service provider to data owner and related data transformation method for preserving business confidential information of the data owner
CN105894336A (zh) * 2016-05-25 2016-08-24 北京比邻弘科科技有限公司 一种基于移动互联网的大数据挖掘方法及系统
CN107480549A (zh) * 2017-06-28 2017-12-15 银江股份有限公司 一种面向数据共享的敏感信息脱敏方法及系统
CN108133294A (zh) * 2018-01-10 2018-06-08 阳光财产保险股份有限公司 基于信息共享的预测方法及装置

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113554476A (zh) * 2020-04-23 2021-10-26 京东数字科技控股有限公司 信用度预测模型的训练方法、系统、电子设备及存储介质
CN113554476B (zh) * 2020-04-23 2024-04-19 京东科技控股股份有限公司 信用度预测模型的训练方法、系统、电子设备及存储介质
CN111915368A (zh) * 2020-07-30 2020-11-10 上海数策软件股份有限公司 汽车行业客户id识别系统、方法及介质
CN111915368B (zh) * 2020-07-30 2024-02-20 上海数策软件股份有限公司 汽车行业客户id识别系统、方法及介质

Also Published As

Publication number Publication date
CN108133294A (zh) 2018-06-08
CN108133294B (zh) 2020-12-04

Similar Documents

Publication Publication Date Title
WO2019137049A1 (zh) 基于信息共享的预测方法、装置、电子设备及计算机存储介质
JP6803980B1 (ja) 信頼されたイニシャライザを用いない秘密分散
CN115943394A (zh) 用于安全纵向联邦学习的方法、装置和系统
Zhao et al. A machine learning based trust evaluation framework for online social networks
CN107305611B (zh) 恶意账号对应的模型建立方法和装置、恶意账号识别的方法和装置
Tang et al. Cloud service QoS prediction via exploiting collaborative filtering and location‐based data smoothing
CN112132676B (zh) 联合训练目标模型的贡献度的确定方法、装置和终端设备
CN110837653B (zh) 标签预测方法、装置以及计算机可读存储介质
CN110166423B (zh) 用户信用的确定方法、装置、系统和数据的处理方法
WO2022174787A1 (zh) 模型训练
CA3117872A1 (en) Clustering techniques for machine learning models
Tariq et al. An analysis of the application of fuzzy logic in cloud computing
CN111800411A (zh) 保护隐私的业务预测模型联合更新方法及装置
JP2016511891A (ja) 大規模データへの妨害攻撃に対するプライバシー
US20230162053A1 (en) Machine-learning techniques for risk assessment based on clustering
CN114398553A (zh) 对象推荐方法、装置、电子设备以及存储介质
Tian et al. Bi-tier differential privacy for precise auction-based people-centric IoT service
CN110838069A (zh) 数据处理方法、装置以及系统
US20220058651A1 (en) Authentication of financial transaction
CN111079153A (zh) 安全建模方法、装置、电子设备及存储介质
CN113989036A (zh) 一种不暴露入模变量的联邦学习预测方法及系统
CN113158047A (zh) 推荐模型训练、信息推送方法、装置、设备及介质
John Joseph et al. A novel trust-scoring system using trustability co-efficient of variation for identification of secure agent platforms
CN111046409B (zh) 一种私有数据多方安全计算方法和系统
Liu et al. A modified TOPSIS method for obtaining the associated weights of the OWA‐type operators

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18899896

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 17/11/2020)

122 Ep: pct application non-entry in european phase

Ref document number: 18899896

Country of ref document: EP

Kind code of ref document: A1