CN114021891A - Living analysis method and device based on federal learning and electronic equipment - Google Patents
Living analysis method and device based on federal learning and electronic equipment Download PDFInfo
- Publication number
- CN114021891A CN114021891A CN202111168116.2A CN202111168116A CN114021891A CN 114021891 A CN114021891 A CN 114021891A CN 202111168116 A CN202111168116 A CN 202111168116A CN 114021891 A CN114021891 A CN 114021891A
- Authority
- CN
- China
- Prior art keywords
- survival
- analysis
- survival analysis
- data
- information
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Abstract
The application provides a survival analysis method and device based on federal learning, electronic equipment and a computer readable storage medium; applied to any one of a plurality of participants, the method comprising: receiving the earliest time and the latest time corresponding to the survival data of other participants, and determining the earliest time and the latest time corresponding to survival analysis according to the earliest time and the latest time corresponding to the survival data of other participants; determining a global time sequence based on the earliest time and the latest time corresponding to the survival analysis; receiving the analysis indexes of each other participant on the medical data, and determining the grouping information of the survival analysis according to the analysis indexes corresponding to each participant; determining first survival analysis information according to the grouping information of the survival analysis and the global time sequence; and receiving second survival analysis information of other multiple participants, and performing survival analysis according to the first survival analysis information and the multiple second survival analysis information.
Description
Technical Field
The present application relates to data processing technologies, and in particular, to a method and an apparatus for survival analysis based on federal learning, an electronic device, and a computer-readable storage medium.
Background
The analysis of survival data composed of data of an event of interest and time data corresponding to the event of interest is becoming more and more important in the fields of biomedicine, sociology, marketology, and the like. When analyzing the survival data, the survival probabilities of the survival data at different time points in a time period are usually calculated first, and then the survival curves are generated according to the survival probabilities corresponding to the different time points respectively.
In the related art, if a plurality of participants analyze the same type of survival data, the survival data of the plurality of participants need to be merged, and then the survival probability and the survival curve corresponding to the merged survival data are calculated. In the process of merging the survival data of the multiple participants, the survival data of the multiple participants are directly exposed to the survival data analysis party, and the survival data of the participants cannot be protected in privacy.
Disclosure of Invention
The embodiment of the application provides a survival analysis method and device based on federated learning, an electronic device and a computer-readable storage medium, which can perform privacy protection on survival data of each participant in a scene of analyzing the survival data of a plurality of participants.
The technical scheme of the embodiment of the application is realized as follows:
in a first aspect, an embodiment of the present application provides a survival analysis method based on federal learning, which is applied to any one of a plurality of participants, and the method includes:
receiving the earliest time and the latest time corresponding to the survival data of other participants, and determining the earliest time and the latest time corresponding to survival analysis according to the earliest time and the latest time corresponding to the survival data of other participants;
determining a global time sequence based on the earliest time and the latest time corresponding to the survival analysis;
receiving the analysis indexes of each other participant on the medical data, and determining the grouping information of the survival analysis according to the analysis indexes corresponding to each participant;
determining first survival analysis information according to the grouping information of the survival analysis and the global time sequence;
and receiving second survival analysis information of other multiple participants, and performing survival analysis according to the first survival analysis information and the multiple second survival analysis information.
In some embodiments, the receiving the earliest time and the latest time corresponding to the survival data of the other participants, and determining the earliest time and the latest time corresponding to the survival analysis according to the earliest time and the latest time corresponding to the survival data of the other participants includes:
receiving the earliest time and the latest time corresponding to the survival data of other multiple participants;
respectively comparing the earliest time corresponding to the local survival data with the earliest times corresponding to the survival data of other multiple participants, and/or respectively comparing the later time corresponding to the local survival data with the latest times corresponding to the survival data of other multiple participants;
and determining the latest moment corresponding to the survival analysis and the latest moment corresponding to the survival analysis according to the comparison result.
In some embodiments, the determining the grouping information of the survival analysis according to the analysis index corresponding to each participant includes:
determining an analysis index corresponding to the local survival data;
determining a global analysis index of survival analysis according to the analysis index of each participant on the medical data and the analysis index corresponding to the local survival data;
and determining the grouping information of the survival analysis according to the global analysis index.
In some embodiments, the determining a global time series based on the earliest and latest moments corresponding to the survival analysis includes:
and determining a global time sequence according to a preset time interval according to the earliest moment and the latest moment corresponding to the survival analysis.
In some embodiments, the performing survival analysis according to the first survival analysis information and a plurality of second survival analysis information includes:
and carrying out operation processing on the first survival analysis information and the plurality of second survival analysis information to obtain global survival analysis information aiming at global analysis indexes.
In some embodiments, the receiving second survival analysis information of other participants and performing survival analysis according to the first survival analysis information and a plurality of second survival analysis information includes:
receiving second survival analysis information which is sent by a plurality of participants and encrypted by using a first function;
encrypting the first biometric analysis information using the first function;
and decrypting the encrypted second survival analysis information and the encrypted first survival analysis information by using a second function to obtain global survival analysis information.
In some embodiments, the performing survival analysis according to the first survival analysis information and a plurality of second survival analysis information includes:
the global survival analysis information comprises at least two survival data groups, and the number of interesting events occurring in each survival data group at each time point in the global time sequence is determined;
determining the survival probability of each survival data group at each time point according to the number of the interesting events;
and constructing a survival curve corresponding to each survival data group based on the survival probability.
In a second aspect, an embodiment of the present application provides a survival analysis device based on federal learning, where the device includes:
determining the earliest time and the latest time corresponding to the survival data of other participants according to the earliest time and the latest time corresponding to the survival data of other participants;
a time sequence determining module, configured to determine a global time sequence based on the earliest time and the latest time corresponding to the survival analysis;
the group information acquisition module is used for receiving the analysis indexes of each other participant on the medical data and determining the group information of the survival analysis according to the analysis indexes corresponding to each participant;
the survival information determining module is used for determining first survival analysis information according to the grouping information of the survival analysis and the global time sequence;
and the survival analysis module is used for receiving second survival analysis information of other multiple participants and carrying out survival analysis according to the first survival analysis information and the multiple second survival analysis information.
In a third aspect, an embodiment of the present application provides an electronic device, including:
a memory for storing executable instructions;
and the processor is used for realizing the survival analysis method based on the federal learning provided by the embodiment of the application when the executable instructions stored in the memory are executed.
In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium storing executable instructions for implementing the federate learning-based survival analysis method provided in the embodiment of the present application when executed by a processor.
According to the survival analysis method based on the federal learning, any one of a plurality of participants receives the earliest time and the latest time corresponding to the survival data of other participants, and the earliest time and the latest time corresponding to the survival analysis are determined according to the earliest time and the latest time corresponding to the survival data of other participants; determining a global time sequence based on the earliest time and the latest time corresponding to the survival analysis; receiving the analysis indexes of each other participant on the medical data, and determining the grouping information of the survival analysis according to the analysis indexes corresponding to each participant; determining first survival analysis information according to the grouping information of the survival analysis and the global time sequence; and receiving second survival analysis information of other multiple participants, and performing survival analysis according to the first survival analysis information and the multiple second survival analysis information. According to the survival analysis method based on federated learning, second survival analysis information of other multiple participants is sent to one participant in an encrypted mode, and privacy of survival data of all participants participating in survival data analysis is guaranteed; each participant only sends the earliest time and the latest time to other participants, and only sends second survival analysis information aiming at the global time sequence to other participants after the global time sequence is determined, and the survival data of each participant is not disclosed, so that the privacy of the data of each participant is protected; the survival analysis is carried out on the basis of the global time sequence constructed at the earliest moment and the latest moment corresponding to the survival analysis, so that the survival data of other participants at moments except the time interval corresponding to the local survival data can be prevented from being missed when the participants analyze the survival data, and the integrity of the survival analysis data is ensured.
Drawings
Fig. 1 is a schematic architecture diagram of a survival data analysis system according to an embodiment of the present application;
fig. 2 is a schematic architecture diagram of a terminal device provided in an embodiment of the present application;
FIG. 3 is a flow chart of a federated learning-based survival analysis method provided in an embodiment of the present application;
FIG. 4 is a schematic diagram of an alternative process for determining a time sequence based on an earliest time and a latest time according to an embodiment of the present application;
fig. 5 is a schematic processing flow diagram for obtaining packet information of target survival data according to an embodiment of the present application;
fig. 6 is a schematic processing flow diagram of survival information corresponding to survival data divided into each survival data group in determination by using a security summation algorithm according to an embodiment of the present application;
FIG. 7 is a schematic diagram of a survival curve 1 provided by an embodiment of the present application;
fig. 8 is a schematic diagram of a survival curve 2 provided in an embodiment of the present application.
Detailed Description
In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be described in further detail with reference to the attached drawings, the described embodiments should not be considered as limiting the present application, and all other embodiments obtained by a person of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.
In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict.
In the following description, references to the terms "first \ second \ third" are only to distinguish similar objects and do not denote a particular order, but rather the terms "first \ second \ third" are used to interchange specific orders or sequences, where appropriate, so as to enable the embodiments of the application described herein to be practiced in other than the order shown or described herein. In the following description, the term "plurality" referred to means at least two.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of the present application only and is not intended to be limiting of the application.
Before further detailed description of the embodiments of the present application, terms and expressions referred to in the embodiments of the present application will be described, and the terms and expressions referred to in the embodiments of the present application will be used for the following explanation.
1) Survival data: the data of the event of interest includes a number of times the event of interest occurred and the time data of the event of interest occurred includes a time point at which the event of interest occurred.
2) And (3) survival data analysis: it is necessary to analyze the event of interest and the time elapsed for the occurrence of the event of interest.
3) Event of interest: events of interest are different for different application domains; in the biomedical field, an event of interest may refer to death of an organism; in the field of sociology, an event of interest may refer to a behavior of a human, such as a countering behavior.
4) Survival probability: refers to the probability that the time T at which the event of interest occurs for an instance is not less than a given time T.
5) Survival curve: the abscissa is time, and the ordinate is survival probability, representing the variation trend of the survival probability in a period of time.
6) Deletion: the absence of any interesting events is referred to as a deletion (censored) for some instances.
Possible reasons for the deletion are:
A. an example is that no event of interest (right-center) is present during the analysis phase.
B. In the analysis phase, the instance is lost.
C. The instance experienced other events than the event of interest resulting in an inability to continue tracking.
6) Example (c): the survival data corresponding to the sample and each instance can use a triple (T)i,Ei,Xi) Represents; wherein, XiFeature vector, T, representing an exampleiIndicating the time of occurrence of the event of interest of the instance.
If the instance has an event of interest, E i1, then TiThe duration from the time point of occurrence of the event to the start time point of the survival data analysis is indicated.
If the instance does not have an event of interest, Ei0, then TiIndicating the time period between the start time point of the survival data analysis and the observation end time point.
Referring to fig. 1, fig. 1 is a schematic structural diagram of a survival data analysis system 100 provided in an embodiment of the present application, a terminal device 400 is connected to a server 200 through a network 300, and the server 200 is connected to a database 500, where the network 300 may be a wide area network or a local area network, or a combination of the two.
In some embodiments, taking an example that one participant is a terminal device, the federate learning-based survival analysis method provided in the embodiments of the present application may be implemented by the terminal device. For example, the terminal device 400 runs the client 410, and the client 410 may be a client for performing survival data analysis.
In some embodiments, taking an example that one participant is a server, the federate learning-based survival analysis method provided in the embodiments of the present application may be cooperatively implemented by the server and the terminal device. For example, the server 200 implementing the federal learning-based survival analysis method receives the earliest time and the latest time corresponding to the survival data of other participants, and determines the earliest time and the latest time corresponding to the survival analysis according to the earliest time and the latest time corresponding to the survival data of the other participants; determining a global time sequence based on the earliest time and the latest time corresponding to the survival analysis; receiving the analysis indexes of each other participant on the medical data, and determining the grouping information of the survival analysis according to the analysis indexes corresponding to each participant; determining first survival analysis information according to the grouping information of the survival analysis and the global time sequence; and receiving second survival analysis information of other multiple participants, and performing survival analysis according to the first survival analysis information and the multiple second survival analysis information. The server 200 sends the results of the survival analysis to the client 410.
In some embodiments, the terminal device 400 or the server 200 may implement the federate learning-based survival analysis method provided in the embodiments of the present application by running a computer program, for example, the computer program may be a native program or a software module in an operating system; or a local (Native) Application program (APP), that is, a program that needs to be installed in an operating system to be executed; or may be an applet, i.e. a program that can be run only by downloading it to the browser environment; but also an applet that can be embedded in any APP. In general, the computer programs described above may be any form of application, module or plug-in.
In some embodiments, the server 200 may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a Cloud server providing basic Cloud computing services such as a Cloud service, a Cloud database, Cloud computing, a Cloud function, Cloud storage, a web service, Cloud communication, a middleware service, a domain name service, a security service, a CDN, and a big data and artificial intelligence platform, where Cloud Technology (Cloud Technology) refers to a hosting Technology for unifying resources of hardware, software, a network, and the like in a wide area network or a local area network to implement computing, storage, processing, and sharing of data. The terminal device 400 may be, but is not limited to, a smart phone, a tablet computer, a notebook computer, a desktop computer, and the like. The terminal device and the server may be directly or indirectly connected through wired or wireless communication, and the embodiment of the present application is not limited.
Taking the example that one participant provided in the embodiment of the present application is a terminal device as an illustration, it can be understood that, for the case that one participant is a server, parts (e.g., a user interface, a presentation module, and an input processing module) in the structure shown in fig. 2 may be absent. Referring to fig. 2, fig. 2 is a schematic structural diagram of a terminal device 400 provided in an embodiment of the present application, where the terminal device 400 shown in fig. 2 includes: at least one processor 460, memory 450, at least one network interface 420, and a user interface 430. The various components in the terminal device 400 are coupled together by a bus system 440. It is understood that the bus system 440 is used to enable communications among the components. The bus system 440 includes a power bus, a control bus, and a status signal bus in addition to a data bus. For clarity of illustration, however, the various buses are labeled as bus system 440 in fig. 2.
The Processor 460 may be an integrated circuit chip having Signal processing capabilities, such as a general purpose Processor, a Digital Signal Processor (DSP), or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, or the like, wherein the general purpose Processor may be a microprocessor or any conventional Processor, or the like.
The user interface 430 includes one or more output devices, including one or more speakers and/or one or more visual display screens, that enable the presentation of media content. The user interface 430 also includes one or more input devices, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.
The memory 450 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid state memory, hard disk drives, optical disk drives, and the like. Memory 450 optionally includes one or more storage devices physically located remote from processor 460.
The memory 450 includes either volatile memory or nonvolatile memory, and may include both volatile and nonvolatile memory. The nonvolatile memory may be a Read Only Memory (ROM), and the volatile memory may be a Random Access Memory (RAM). The memory 450 described in embodiments herein is intended to comprise any suitable type of memory.
In some embodiments, memory 450 is capable of storing data, examples of which include programs, modules, and data structures, or a subset or superset thereof, to support various operations, as exemplified below.
An operating system 451, including system programs for handling various basic system services and performing hardware-related tasks, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and handling hardware-based tasks;
a network communication module 452 for communicating to other computing devices via one or more (wired or wireless) network interfaces 420, exemplary network interfaces 420 including: bluetooth, wireless compatibility authentication (WiFi), and Universal Serial Bus (USB), etc.;
a rendering module 453 for enabling the rendering of information (e.g., user interfaces for operating peripherals and displaying content and information) via one or more output devices (e.g., display screens, speakers, etc.) associated with user interface 430;
an input processing module 454 for medical semantic mining of one or more user inputs or interactions from one of the one or more input devices.
In some embodiments, the apparatus provided in the embodiments of the present application may be implemented in software, and fig. 2 illustrates the survival data analysis apparatus 455 stored in the memory 450, which may be software in the form of programs and plug-ins, and may include the following software modules: an acquisition module 4551, a time-series determination module 4552, a grouping information acquisition module 4553, an existence information determination module 4554, and an existence analysis module 4555, which are logical and thus may be arbitrarily combined or further divided according to the functions implemented. The functions of the respective modules will be explained below.
The survival analysis method based on federal learning provided by the embodiment of the application will be described below with reference to an exemplary application and implementation of a participant provided by the embodiment of the application.
Referring to fig. 3, fig. 3 is a schematic flow chart of an alternative method for survival analysis based on federal learning according to an embodiment of the present application, which will be described with reference to the steps shown in fig. 3.
And S101, receiving the earliest time and the latest time corresponding to the survival data of other participants, and determining the earliest time and the latest time corresponding to the survival analysis according to the earliest time and the latest time corresponding to the survival data of other participants.
In some embodiments, where each participant corresponds to a piece of survival data, the survival data may be medical data for the medical domain. Each of the survival data includes data of an event of interest including a number of times the event of interest occurred and time data of when the event of interest occurred including a time point at which the event of interest occurred. Survival data may be as shown in table 1 below: the "T" column indicates the time when the event occurred, and the "E" column indicates the event of interest; wherein the value of the "E" column is "1", indicating the occurrence of a death event; the value of the "E" column is "0", indicating that a deletion event has occurred. As can be seen from table 1, at time point 2, 1 death event and 1 deletion event occurred. At time point 4, 1 erasure event occurred. At time point 5, 2 death events occurred. At time point 8, 3 erasure events occurred.
TABLE 1 survival data
| | E | |
| 2 | 1 | |
| 2 | 0 | |
| 4 | 0 | |
| 5 | 1 | |
| 5 | 1 | |
| 8 | 0 | |
| 8 | 0 | |
| 8 | 0 |
In some embodiments, the earliest and latest times corresponding to the survival analysis may refer to the earliest and latest times of the survival data corresponding to all participants. As an example, if there are three participants, the survival data corresponding to all the participants includes three pieces of survival data, namely survival data a, survival data B and survival data C; the survival data a, the survival data B, and the survival data C constitute a survival data set. The earliest time corresponding to the survival data A is 24 points at 1 month and 1 day in 2000, and the latest time corresponding to the survival data A is 24 points at 1 month and 3 months in 2000; the earliest moment corresponding to the survival data B is 24 points in 2 months and 1 day in 2000, and the latest moment corresponding to the survival data A is 24 points in 4 months and 1 day in 2000; the earliest moment corresponding to the survival data C is 24 points in 3 months and 1 day in 2000, and the latest moment corresponding to the survival data C is 24 points in 5 months and 1 day in 2000; the earliest time corresponding to the target survival data is 24 points 1/2000, and the latest time corresponding to the survival analysis is 24 points 1/5/2000.
In some embodiments, the earliest and latest times corresponding to the survival analysis may be determined using a safety comparison algorithm.
In some embodiments, the participant who implements the federate learning-based survival analysis method receives the earliest time and the latest time corresponding to the survival data of other participants, and determines the earliest time and the latest time corresponding to the survival analysis according to the earliest time and the latest time corresponding to the survival data of other participants can be implemented by the following technical solutions:
receiving the earliest time and the latest time corresponding to the survival data of other multiple participants; respectively comparing the earliest time corresponding to the local survival data with the earliest times corresponding to the survival data of other multiple participants, and/or respectively comparing the latest time corresponding to the local survival data with the latest times corresponding to the survival data of other multiple participants; and determining the latest moment corresponding to the survival analysis and the latest moment corresponding to the survival analysis according to the comparison result.
In specific implementation, one participant and the other participant compare the size of the earliest moment corresponding to the survival data by using a safety comparison algorithm; the participant corresponding to the smaller of the two earliest moments and the third participant compare the sizes of the earliest moments corresponding to the respective survival data by using the safety comparison algorithm; and repeating the steps until the earliest moment corresponding to the target survival data is determined.
As an example, one participant compares the size of the earliest time corresponding to the two participants' survival data with the other participant using a secure comparison algorithm. And if the earliest moment corresponding to the survival data corresponding to one participant is smaller than the size of the earliest moment corresponding to the survival data corresponding to the other participant, comparing the sizes of the earliest moments corresponding to the survival data of the two participants by the one participant and the third participant with smaller earliest moments by using a safety comparison algorithm. If the earliest moment corresponding to the survival data corresponding to the third participant is smaller than the size of the earliest moment corresponding to the survival data corresponding to one participant, the third participant and the fourth participant compare the sizes of the earliest moments corresponding to the survival data of the two participants by using a safety comparison algorithm; and repeating the steps until the earliest moment corresponding to the survival analysis is determined.
In some embodiments, the security comparison algorithm includes multiple implementation forms, and the following description will describe the size of the earliest time when one participant compares the survival data corresponding to each other with another participant by taking the security comparison algorithm as an example that uses Paillier homomorphic encryption. The earliest time corresponding to the survival data corresponding to one participant is a, and the earliest time corresponding to the survival data corresponding to the other participant is b. One participant generates a Pai llier key, a public key PK and a private key SK, one participant encrypts a by using the public key PK to obtain E (a), and E (a) and the public key PK are sent to the other participant. The other party selects random numbers x and y, calculates the value of b x + y, and encrypts b x + y by using a public key PK to obtain E (b x + y); the other participant calculates E (a × x + y) ═ E (a) × x + E (y) using E (a) under homomorphic encryption, and sends E (b × x + y) and E (a × x + y) to the one participant. One participant decrypts E (b x + y) and E (a x + y) by using the private key SK to obtain b x + y and a x + y; comparing the sizes of b x + y and a x + y to obtain the size relationship between a and b; and one participant sends the calculated magnitude relation of a and b to the other participant. Therefore, one participant and the other participant can know the sizes of the a and the b, and the one participant and the other participant do not need to know survival data of the other participant, so that the privacy of the survival data of the one participant and the other participant is ensured.
In some embodiments, the latest time corresponding to the target survival data acquired by one participant may be implemented by the following technical solutions:
comparing the size of the latest moment corresponding to the survival data of one participant and any one participant (a fourth participant) in a plurality of participants by using a safety comparison algorithm; comparing the participant corresponding to the larger of the two latest moments with other participants (a fifth participant) in the plurality of participants by using the safety comparison algorithm, wherein the participant corresponds to the largest of the two latest moments; and repeating the steps until the latest moment corresponding to the survival analysis is determined.
As an example, one participant and a fourth participant compare the magnitude of their respective latest time instants using a secure comparison algorithm. If the latest time corresponding to the survival data corresponding to one participant is smaller than the latest time corresponding to the survival data corresponding to the fourth participant, the latest time corresponding to the survival data corresponding to the fourth participant and the other participants (the fifth participant) in the plurality of participants are compared by using a safety comparison algorithm. If the latest moment corresponding to the survival data corresponding to the fifth party is larger than the latest moment corresponding to the survival data corresponding to the fourth party, comparing the latest moment corresponding to the survival data corresponding to the fifth party with the latest moment corresponding to the survival data corresponding to the sixth party by the fifth party and the sixth party by using a safety comparison algorithm; and repeating the steps until the latest moment corresponding to the target survival data is determined.
In some embodiments, the security comparison algorithm that determines the latest moment corresponding to the survival analysis may also include multiple implementation manners, such as a Paillier homomorphic encryption manner adopted by the security comparison algorithm that determines the earliest moment corresponding to the survival analysis, or other algorithms.
In some embodiments, if the survival data corresponding to the earliest time and/or the survival data corresponding to the latest time do not belong to a participant implementing the federal learning-based survival analysis method, the participant implementing the federal learning-based survival analysis method receives first broadcast information, where the first broadcast information carries the earliest time and/or the latest time of the survival analysis. If the survival data corresponding to the earliest moment and/or the survival data corresponding to the latest moment belong to a first participant, the participant implementing the survival analysis method based on the federal learning sends first broadcast information, and the first broadcast information carries the earliest moment and/or the latest moment of the survival analysis, so that other participants participating in the survival analysis can know the earliest moment and the latest moment.
It should be noted that the manner of determining the earliest time and the latest time of the survival analysis based on the security comparison algorithm is applicable to the case where the number of participants is three or more.
And step S102, determining a global time sequence based on the earliest time and the latest time corresponding to the survival analysis.
In some embodiments, an optional process flow for determining the global time series based on the earliest and latest moments corresponding to the survival analysis, as shown in fig. 4, may include at least the following steps:
step S102a, determining that the global time sequence includes a time sub-sequence and at least one time point.
Step S102b, determining a starting time of the time sub-sequence as an earliest time, and a time interval between adjacent time points in the time sub-sequence as a target time interval.
Step S102c, determining that the at least one time point includes a latest time instant, the latest time instant being located after the time sub-sequence.
In some embodiments, for the step S102b, if the earliest time corresponding to the survival analysis is TminThe latest time corresponding to the survival analysis is TmaxTarget time interval of TintervalThen the time subsequence sequentially comprises: t ismin、Tmin+Tinterval、Tmin+2Tinterval…Tmin+nTinterval(ii) a And Tmin+nTintervalLess than Tmax。
In some embodiments, for step S102c above, at least one time point is a latest time T corresponding to the storage analysismaxThen, the global time sequence sequentially includes: t ismin、Tmin+Tinterval、Tmin+2Tinterval…Tmin+nTinterval、Tmax。
In other embodiments, for step S102c above, the at least one time point includes: latest time T corresponding to standard survival datamaxAnd Tmax+TintervalThen the time sequence sequentially includes: t ismin、Tmin+Tinterval、Tmin+2Tinterval…Tmin+nTinterval、Tmax、Tmax+Tinterval。
Step S103, receiving the analysis indexes of each other participant to the medical data, and determining the grouping information of the survival analysis according to the analysis indexes corresponding to each participant.
In some embodiments, one participant receives second broadcast information sent by other participants, and the second broadcast information carries analysis indexes of medical data of other participants.
Wherein the analysis index of the medical data by each participant is determined based on the first characteristic. The first characteristic may be a characteristic for grouping the memory data. As an example, the analysis of the survival data corresponding to the second of the other participants may include using drug a and using drug C; the analysis of the survival data for a third of the other participants may include using drug B and using drug C. The analytical index of each participant for the medical data includes use of drug a, use of drug B, and use of drug C.
In some embodiments, a process flow of determining grouping information of survival analysis according to the analysis index corresponding to each participant may be as shown in fig. 5, and includes the following steps:
step S103a, determining an analysis index corresponding to the local survival data.
In some embodiments, a participant groups local survival data based on the first characteristic to obtain an analysis index of the local survival data. Where the local survival data may be a participant in a survival analysis method that performs federated learning. The first characteristic may be a characteristic for grouping the memory data. As an example, the first characteristic may be the kind of medication used, and the local survival data includes: and using the survival data sets corresponding to the medicine A and the medicine B, wherein the analysis indexes corresponding to the local survival data comprise the use of the medicine A and the use of the medicine B. As an example, the first characteristic may also be whether or not an interventional procedure is used, the kind of surgery, and the like.
In some embodiments, the participant performing the federal learning based survival analysis method may also broadcast the first packet information to other participants participating in the survival analysis.
Step S103b, determining grouping information of survival analysis according to the analysis index of each participant on the medical data and the analysis index corresponding to the local survival data.
In some embodiments, a participant performing a federal learning based survival analysis method may receive grouping information of other participants, determine an analysis index for medical data for each participant; the union of the analysis indexes of each participant on the medical data and the analysis indexes corresponding to the local survival data is a global analysis index; and determining the grouping information of the survival analysis according to the global survival index. As an example, the analysis indexes corresponding to the local survival data include a medicine a and a medicine B, the analysis indexes of the participant on the medical data include a medicine a, a medicine B and a medicine C, and the grouping information of the survival analysis includes: the survival data of all participants uses the survival data group of the drug A, the survival data of all participants uses the survival data group of the drug B, and the survival data of all participants uses the survival data group of the drug C.
And step S104, determining first survival analysis information according to the grouping information of the survival analysis and the global time sequence.
In some embodiments, the global time sequence is used as a time axis, and the local survival data is grouped according to the grouping information of the survival analysis to obtain first survival analysis information corresponding to the local survival data.
In some embodiments, in the medical field, the first survival analysis information may refer to a number of people whose local survival data survived at a time. Alternatively, the number of people who have had an event of interest and the number of people who have been lost may be calculated first; and subtracting the number of people with the interesting event and the number of lost people from the total number of people with the local survival data to obtain the number of people living at a certain moment. The number of people information table corresponding to the local survival data can be as shown in the following table 2: t represents time, dn represents the number of people who have an event of interest (updated atTn) at the corresponding time, and the number of people who have an event of interest can be the number of people who have died. The number of people diverted (removed at Tn) includes the sum of the number of people who occurred the event of interest and the number of people deleted. The number of surviving people Yn (at risk at Tn) may also be referred to as the number of people at risk. If the local survival data are shown in the following table 2, T represents a time point, E ═ 1 represents a death event, and E ═ 0 represents a deletion event; the first biometric information may refer to Yn in table 3 below.
TABLE 2 local survival data
| | E | |
| 2 | 1 | |
| 2 | 0 | |
| 4 | 0 | |
| 5 | 1 | |
| 5 | 1 | |
| 8 | 0 | |
| 8 | 0 | |
| 8 | 0 |
TABLE 3 people number information table corresponding to local survival data
Step S105, receiving second survival analysis information of other multiple participants, and performing survival analysis according to the first survival analysis information and multiple second survival analysis information.
In some embodiments, the process of performing survival analysis according to the first survival analysis information and a plurality of second survival analysis information may include the following steps as shown in fig. 6:
step S105a is to receive the second survival analysis information encrypted by the first function and transmitted by the plurality of participants.
In some embodiments, the other participant encrypts the second liveness analysis information using the first function; and sending the encrypted second survival analysis information to a participant implementing a federal learning-based survival analysis method. Wherein one second survival analysis information corresponds to survival data of one participant. The encrypted second survival analysis information may be q2(xi),...,qn(xi). Each participant separately targets all xjE.g. X, calculate qi(xj) And sends qi(xj) To the other participants.
In some embodiments, the first function may be an n-1 th order polynomial as follows:
qi(X)＝aiXn-1+biXn-2+...+vi；
wherein v isiMay be survival information, aiAnd biIs a coefficient of a first function;
step S105b, encrypting the first biometric analysis information using the first function.
In some embodiments, the first biometric analysis information is encrypted using a first function. Multiple participants share a set of random values X ═ X1,x2,...,xnIf a participant utilizes a first function pair v1Encrypting to obtain a first value q1(xj). Each participant enjoys a private value viIn the embodiment of the present application, the private value is Yn in table 2.
Step S105c, decrypting the encrypted second survival analysis information and the encrypted first survival analysis information by using a second function, to obtain global survival analysis information.
In some embodiments, the first electronic device pairs q by the following formula1(xi),q2(xi),...,qn(xi) The summation is performed.
resulti＝q1(xi)+q2(xi)+...+qn(xi)
＝(a1+a2+...+an)xi n-1+(b1+b2+...+bn)xi n-2+...+(v1+v2+...+vn)
Each participant can pair q1(xi),q2(xi),...,qn(xi) Summing to obtain equation set, and storingiTo other participants.
resulti＝Axi n-1+Bxi n-2+...+Vsum，1≤i≤n，A＝a1+a2+...+an，B＝b1+b2+...+bn(ii) a The system of equations can also be called as a second function, and survival information (using V) corresponding to the target survival data can be obtained by solving the system of equationssumIs represented by V), Vsum＝v1+v2+...+vn(ii) a Wherein A ═ a1+a2+...+an，B＝b1+b2+...+bn，1≤i≤n。
Therefore, each participant can obtain the survival information corresponding to other participants without knowing the survival data of other participants, and the privacy of the survival data among the participants is ensured.
The algorithm adopted in steps S105a to S105c is a secure sum algorithm, and the secure sum algorithm has multiple implementation forms, which is not limited in the embodiment of the present application.
In some embodiments, the global survival analysis information includes at least two survival data sets, and the implementing of the survival analysis according to the first survival analysis information and the plurality of second survival analysis information may include: determining the number of interesting events occurring in each survival data group at each time point in the global time sequence; determining the survival probability of each survival data group at each time point according to the number of the interesting events; and constructing a survival curve corresponding to each survival data group based on the survival probability.
In some embodiments, for the nth time point, the survival probability may be calculated by:
Wherein, S (t)n-1) Refers to the probability of survival at time t (n-1), dn refers to the number of events that occurred at time tn (observed to death), and Yn refers to the person who still survived as soon as time tn passes (at risk) (if there is an instance deletion between t (n-1) and tn, then the deleted patient should be removed from the total survival data when calculating Yn. Meanwhile, tn is 0, and S (0) is 1.
In some embodiments, a continuous curve is drawn according to the corresponding relationship between each time point in the actual sequence and the survival probability of each survival data group at the time point, that is, the survival curve is obtained.
And step S106, determining the difference between the survival curves corresponding to at least two survival data groups by using a target detection mode.
In some embodiments, the target verification method may be Logrank verification, Cox model verification, or the like.
In some embodiments, each survival data group corresponds to a survival curve, and each survival data group corresponds to a characteristic factor, so that differences inevitably exist between the survival curves of different survival data groups corresponding to different characteristic factors, and the influence of the characteristic factors on the survival data can be judged based on the differences between the survival curves.
In some embodiments, the process of determining the difference between the survival curves based on the Logrank test mode is the same as the prior art, and is not described herein again.
The survival analysis method based on federal learning provided in the embodiments of the present application is described below by taking the case where the survival analysis method based on federal learning is applied to the biomedical field.
The earliest moment of obtaining the target survival data by using a safety comparison algorithm is 2, and the latest moment is 8; according to the broadcast information sent by each participant, whether the characteristic factors intervene in treatment can be known; survival data corresponding to each participant are divided into two groups aiming at characteristic factors, namely interventional therapy and non-interventional therapy.
The survival information of each participant for obtaining the target survival data by using the safety summation algorithm is shown in the following tables 4 and 5: x1 denotes a feature vector; if X1 ═ 0 represents no intervention, and X1 ═ 1 represents intervention.
TABLE 4 survival information (1) (X)1＝0)
| T | dn:observed at Tn | removed at Tn | Yn:at_risk at |
| 2 | 1 | 1 | 3 |
| 5 | 1 | 1 | 2 |
| 8 | 0 | 1 | 1 |
TABLE 5 survival information (2) (X)1＝1)
The survival curve 1 plotted based on table 4 is shown in fig. 7, and the survival curve 2 plotted based on table 5 is shown in fig. 8. The difference between the survival curves 1 and 2 was then examined using Logrank. Wherein different electronic devices can check the difference between the survival curves 1 and 2 using different checking methods.
According to the survival analysis method based on the federal learning, any one of a plurality of participants receives the earliest time and the latest time corresponding to the survival data of other participants, and the earliest time and the latest time corresponding to the survival analysis are determined according to the earliest time and the latest time corresponding to the survival data of other participants; determining a global time sequence based on the earliest time and the latest time corresponding to the survival analysis; receiving the analysis indexes of each other participant on the medical data, and determining the grouping information of the survival analysis according to the analysis indexes corresponding to each participant; determining first survival analysis information according to the grouping information of the survival analysis and the global time sequence; and receiving second survival analysis information of other multiple participants, and performing survival analysis according to the first survival analysis information and the multiple second survival analysis information. By obtaining the earliest time and the latest time corresponding to the survival analysis, the timestamps of the survival data analyzed by all the participants can be unified, and the missing of the survival data is avoided. Because the earliest moment and the latest moment corresponding to the survival analysis are determined by using the safety comparison algorithm, each participant can not expose own survival data, and the privacy of the survival data is ensured. The survival information corresponding to each survival data group is determined by utilizing a safety summation algorithm, so that each participant can be prevented from exposing the self survival data, and the privacy of the survival data is ensured. Each participant only sends the earliest time and the latest time to other participants, and after the global time sequence is determined, only sends second survival analysis information aiming at the global time sequence to other participants without revealing the survival data of the participant, so that the privacy of the data of each participant is protected.
Continuing with the exemplary structure of the federate learning based survival analysis device 455 provided in this embodiment of the present application implemented as software modules, in some embodiments, as shown in fig. 2, the software modules stored in the survival data analysis device 455 of the memory 450 may include: an obtaining module 4551, configured to determine the earliest time and the latest time corresponding to the survival data of the other participants, according to the earliest time and the latest time corresponding to the survival data of the other participants; a time sequence determining module 4552, configured to determine a global time sequence based on the earliest time and the latest time corresponding to the survival analysis; the grouping information obtaining module 4553 is configured to receive an analysis index of each of the other participants for the medical data, and determine grouping information of survival analysis according to the analysis index corresponding to each of the participants; a survival information determining module 4554, configured to determine first survival analysis information according to the grouping information of survival analysis and the global time series; and a survival analysis module 4555, configured to receive second survival analysis information of the other multiple participants, and perform survival analysis according to the first survival analysis information and the multiple second survival analysis information.
In some embodiments, the obtaining module 4551 is configured to receive the earliest time and the latest time corresponding to the survival data of the other multiple participants;
respectively comparing the earliest time corresponding to the local survival data with the earliest times corresponding to the survival data of other multiple participants, and/or respectively comparing the later time corresponding to the local survival data with the latest times corresponding to the survival data of other multiple participants;
and determining the latest moment corresponding to the survival analysis and the latest moment corresponding to the survival analysis according to the comparison result.
In some embodiments, the grouping information obtaining module 4553 is configured to determine an analysis indicator corresponding to local survival data;
and determining grouping information of survival analysis according to the analysis index of each participant on the medical data and the analysis index corresponding to the local survival data.
In some embodiments, the time sequence determining module 4552 is configured to determine the global time sequence according to a preset time interval according to the earliest time and the latest time corresponding to the survival analysis.
In some embodiments, the survival analysis module 4555 is configured to perform operation processing on the first survival analysis information and the plurality of second survival analysis information to obtain global survival analysis information for a global analysis indicator.
In some embodiments, the survival analysis module 4555 is configured to receive second survival analysis information sent by a plurality of participants and encrypted by using a first function;
encrypting the first biometric analysis information using the first function;
and decrypting the encrypted second survival analysis information and the encrypted first survival analysis information by using a second function to obtain global survival analysis information.
In some embodiments, the survival analysis module 4555, wherein the global survival analysis information includes at least two survival data sets, and is configured to determine the number of events of interest occurring in each of the survival data sets at each time point in the global time series;
determining the survival probability of each survival data group at each time point according to the number of the interesting events;
and constructing a survival curve corresponding to each survival data group based on the survival probability.
Embodiments of the present application provide a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, so that the computer device executes the federal learning based survival analysis method described in this embodiment.
Embodiments of the present application provide a computer-readable storage medium having stored therein executable instructions that, when executed by a processor, cause the processor to perform a method provided by embodiments of the present application, for example, a federal learning based survival analysis method as shown in fig. 3 to 8.
In some embodiments, the computer-readable storage medium may be memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash, magnetic surface memory, optical disk, or CD-ROM; or may be various devices including one or any combination of the above memories.
In some embodiments, executable instructions may be written in any form of programming language (including compiled or interpreted languages), in the form of programs, software modules, scripts or code, and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.
By way of example, executable instructions may correspond, but do not necessarily have to correspond, to files in a file system, and may be stored in a portion of a file that holds other programs or data, such as in one or more scripts in a HyperText Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files that store one or more modules, sub-programs, or portions of code).
By way of example, executable instructions may be deployed to be executed on one computing device or on multiple computing devices at one site or distributed across multiple sites and interconnected by a communication network.
The above description is only an example of the present application, and is not intended to limit the scope of the present application. Any modification, equivalent replacement, and improvement made within the spirit and scope of the present application are included in the protection scope of the present application.
Claims (10)
1. A federal learning-based survival analysis method applied to any one of a plurality of participants, the method comprising:
receiving the earliest time and the latest time corresponding to the survival data of other participants, and determining the earliest time and the latest time corresponding to survival analysis according to the earliest time and the latest time corresponding to the survival data of other participants;
determining a global time sequence based on the earliest time and the latest time corresponding to the survival analysis;
receiving the analysis indexes of each other participant on the medical data, and determining the grouping information of the survival analysis according to the analysis indexes corresponding to each participant;
determining first survival analysis information according to the grouping information of the survival analysis and the global time sequence;
and receiving second survival analysis information of other multiple participants, and performing survival analysis according to the first survival analysis information and the multiple second survival analysis information.
2. The method of claim 1, wherein receiving the earliest and latest moments corresponding to the survival data of the other participants and determining the earliest and latest moments corresponding to the survival analysis according to the earliest and latest moments corresponding to the survival data of the other participants comprises:
receiving the earliest time and the latest time corresponding to the survival data of other multiple participants;
respectively comparing the earliest time corresponding to the local survival data with the earliest times corresponding to the survival data of other multiple participants, and/or respectively comparing the later time corresponding to the local survival data with the latest times corresponding to the survival data of other multiple participants;
and determining the latest moment corresponding to the survival analysis and the latest moment corresponding to the survival analysis according to the comparison result.
3. The method of claim 1, wherein determining the grouping information of the survival analysis according to the analysis index corresponding to each participant comprises:
determining an analysis index corresponding to the local survival data;
and determining grouping information of survival analysis according to the analysis index of each participant on the medical data and the analysis index corresponding to the local survival data.
4. The method of claim 1, wherein determining a global time series based on the earliest and latest moments corresponding to the survival analysis comprises:
and determining a global time sequence according to a preset time interval according to the earliest moment and the latest moment corresponding to the survival analysis.
5. The method of claim 1, wherein performing a survival analysis based on the first survival analysis information and a plurality of second survival analysis information comprises:
and carrying out operation processing on the first survival analysis information and the plurality of second survival analysis information to obtain global survival analysis information aiming at global analysis indexes.
6. The method of claim 1, wherein receiving second survival analysis information from a plurality of other participants and performing survival analysis based on the first survival analysis information and a plurality of the second survival analysis information comprises:
receiving second survival analysis information which is sent by a plurality of participants and encrypted by using a first function;
encrypting the first biometric analysis information using the first function;
and decrypting the encrypted second survival analysis information and the encrypted first survival analysis information by using a second function to obtain global survival analysis information.
7. The method according to claim 5 or 6, wherein performing survival analysis according to the first survival analysis information and a plurality of second survival analysis information comprises:
the global survival analysis information comprises at least two survival data groups, and the number of interesting events occurring in each survival data group at each time point in the global time sequence is determined;
determining the survival probability of each survival data group at each time point according to the number of the interesting events;
and constructing a survival curve corresponding to each survival data group based on the survival probability.
8. A federal learning-based survival analysis device, the device comprising:
the acquisition module is used for determining the earliest time and the latest time corresponding to the survival data of other participants according to the earliest time and the latest time corresponding to the survival data of other participants;
a time sequence determining module, configured to determine a global time sequence based on the earliest time and the latest time corresponding to the survival analysis;
the group information acquisition module is used for receiving the analysis indexes of each other participant on the medical data and determining the group information of the survival analysis according to the analysis indexes corresponding to each participant;
the survival information determining module is used for determining first survival analysis information according to the grouping information of the survival analysis and the global time sequence;
and the survival analysis module is used for receiving second survival analysis information of other multiple participants and carrying out survival analysis according to the first survival analysis information and the multiple second survival analysis information.
9. An electronic device, comprising:
a memory for storing executable instructions;
a processor configured to implement the federated learning-based survival analysis method of any of claims 1 to 7 when executing the executable instructions stored in the memory.
10. A computer readable storage medium having stored thereon executable instructions for executing, by a processor, the method for federated learning-based survival analysis of any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111168116.2A CN114021891B (en) | 2021-09-30 | 2021-09-30 | Living analysis method and device based on federal learning and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111168116.2A CN114021891B (en) | 2021-09-30 | 2021-09-30 | Living analysis method and device based on federal learning and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114021891A true CN114021891A (en) | 2022-02-08 |
| CN114021891B CN114021891B (en) | 2022-10-14 |
Family
ID=80055331
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111168116.2A Active CN114021891B (en) | 2021-09-30 | 2021-09-30 | Living analysis method and device based on federal learning and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114021891B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180218171A1 (en) * | 2017-01-31 | 2018-08-02 | Hewlett Packard Enterprise Development Lp | Performing privacy-preserving multi-party analytics on horizontally partitioned local data |
| CN109285344A (en) * | 2018-06-25 | 2019-01-29 | 江苏智通交通科技有限公司 | The emphasis monitoring object recognition methods of traffic high-risk personnel and intelligent decision system |
| CN110084455A (en) * | 2018-01-26 | 2019-08-02 | 阿里巴巴集团控股有限公司 | A kind of data processing method, apparatus and system |
| CN110459276A (en) * | 2019-08-15 | 2019-11-15 | 北京嘉和海森健康科技有限公司 | A kind of data processing method and relevant device |
| US20200065713A1 (en) * | 2018-08-24 | 2020-02-27 | Adobe Inc. | Survival Analysis Based Classification Systems for Predicting User Actions |
| CN111243738A (en) * | 2019-12-31 | 2020-06-05 | 东软集团股份有限公司 | Survival analysis model construction method, survival rate prediction method, device and equipment |
| CN111913999A (en) * | 2020-06-08 | 2020-11-10 | 华南理工大学 | Statistical analysis method, system and storage medium based on multiomic and clinical data |
| WO2021008017A1 (en) * | 2019-07-17 | 2021-01-21 | 深圳前海微众银行股份有限公司 | Federation learning method, system, terminal device and storage medium |
-
2021
- 2021-09-30 CN CN202111168116.2A patent/CN114021891B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180218171A1 (en) * | 2017-01-31 | 2018-08-02 | Hewlett Packard Enterprise Development Lp | Performing privacy-preserving multi-party analytics on horizontally partitioned local data |
| CN110084455A (en) * | 2018-01-26 | 2019-08-02 | 阿里巴巴集团控股有限公司 | A kind of data processing method, apparatus and system |
| CN109285344A (en) * | 2018-06-25 | 2019-01-29 | 江苏智通交通科技有限公司 | The emphasis monitoring object recognition methods of traffic high-risk personnel and intelligent decision system |
| US20200065713A1 (en) * | 2018-08-24 | 2020-02-27 | Adobe Inc. | Survival Analysis Based Classification Systems for Predicting User Actions |
| WO2021008017A1 (en) * | 2019-07-17 | 2021-01-21 | 深圳前海微众银行股份有限公司 | Federation learning method, system, terminal device and storage medium |
| CN110459276A (en) * | 2019-08-15 | 2019-11-15 | 北京嘉和海森健康科技有限公司 | A kind of data processing method and relevant device |
| CN111243738A (en) * | 2019-12-31 | 2020-06-05 | 东软集团股份有限公司 | Survival analysis model construction method, survival rate prediction method, device and equipment |
| CN111913999A (en) * | 2020-06-08 | 2020-11-10 | 华南理工大学 | Statistical analysis method, system and storage medium based on multiomic and clinical data |
Non-Patent Citations (3)
| Title |
|---|
| ROBERTS MICHAEL ET AL.: "Common pitfalls and recommendations for using machine learning to detect and prognosticate for COVID-19 using chest radiographs and CT scans", 《NATURE MACHINE INTELLIGENCE》 * |
| 李杏: "生存分析的基本方法及其运用", 《华南预防医学》 * |
| 窦佐超 等: "联盟学习在生物医学大数据隐私保护中的原理与应用", 《医学信息学杂志》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114021891B (en) | 2022-10-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7089014B2 (en) | Systems and methods for anonymizing health data and modifying and editing health data across geographic areas for analysis | |
| Kuzu et al. | A practical approach to achieve private medical record linkage in light of public resources | |
| US20180046766A1 (en) | System for rapid tracking of genetic and biomedical information using a distributed cryptographic hash ledger | |
| Sharma et al. | Design of blockchain-based precision health-care using soft systems methodology | |
| Rajendran et al. | Cloud-based federated learning implementation across medical centers | |
| US20130179176A1 (en) | Computer implemented method for determining the presence of a disease in a patient | |
| Kim et al. | Sharing medical questionnaries based on blockchain | |
| EP3534287A1 (en) | Inserting a further data block into a first ledger | |
| Sufi et al. | Tracking anti-vax social movement using AI-based social media monitoring | |
| Marshall | The benefit of early survival on PD versus HD—Why this is (still) very important | |
| Denecke | An ethical assessment model for digital disease detection technologies | |
| Bonofiglio et al. | Recovery of original individual person data (IPD) inferences from empirical IPD summaries only: applications to distributed computing under disclosure constraints | |
| Ismail et al. | Performance evaluation of a patient-centric blockchain-based healthcare records management framework | |
| WO2021030226A1 (en) | Verifying data accuracy in privacy-preserving computations | |
| CN111986763A (en) | Disease data analysis method and device, electronic device and storage medium | |
| O'Keefe et al. | Confidentialising survival analysis output in a remote data access system | |
| US10025840B2 (en) | Method and system for making multisite performance measure anonymous and for controlling actions and re-identification of anonymous data | |
| CN114021891B (en) | Living analysis method and device based on federal learning and electronic equipment | |
| CN111814169A (en) | Digestive tract disease data encryption obtaining method and risk prediction system | |
| US11694774B1 (en) | Platform for perpetual clinical collaboration and innovation with patient communication using anonymized electronic health record data, clinical, and patient reported outcomes and data | |
| Alfaidi et al. | Health Record Chain (HRC): Implementation of Mobile Healthcare system using Blockchain to enhance Privacy of Electronic Health Record EHR | |
| Babenko et al. | Development of the algorithm to ensure the protection of confidential data in cloud medical information system | |
| Gautam et al. | Blockchain for secure and proper management of medical data and records | |
| US10755812B1 (en) | System and method for timely notification of treatment | |
| Cong et al. | A Trustworthy Decentralized System for Health Data Integration and Sharing: Design and Experimental Validation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |