CN-121979943-A - Cross-department data collaborative analysis method and system integrating federal learning

Abstract

The invention provides a cross-department data collaborative analysis method and a system for merging federal learning, which relate to the technical field of federal learning, and are characterized in that a cross-department federal learning collaborative link is dynamically constructed and a configuration scheme is generated; the method comprises the steps of providing federal collaborative capability for local data processing units of all departments based on a configuration scheme, generating federal energized local data processing units, generating cross-department federal collaborative data link flows through interaction with federal learning collaborative centers, constructing a cross-department data collaborative analysis model based on the cross-department federal collaborative data link flows, and completing cross-department data collaborative analysis and outputting results by using the model. The invention effectively solves the problems of privacy protection and safety compliance in cross-department data collaboration, and realizes the efficient fusion and deep analysis of data.

Inventors

• LIN CANFENG
• WENG SHILIANG
• LI YITAN
• SHI ZECONG

Assignees

• 广州乐税信息科技有限公司

Dates

Publication Date

20260505

Application Date

20260326

Claims (10)

1. 1. A cross-department data collaborative analysis method fusing federal learning, the method comprising: Dynamically constructing a cross-department federation learning cooperative link, wherein the cross-department federation learning cooperative link is used for associating local data processing units of all departments with federation learning cooperative centers to generate a cross-department federation learning cooperative link configuration scheme; The federal collaborative capability is given to the local data processing units of each department based on the cross-department federal learning collaborative link configuration scheme, and federal energized local data processing units of each department are generated; Generating cross-department federation cooperative data link flows through interaction between federation energized local data processing units of all departments and federation learning cooperative centers; and constructing a cross-department data collaborative analysis model based on the cross-department federation collaborative data link flow, completing cross-department data collaborative analysis by using the cross-department data collaborative analysis model, and outputting a cross-department data collaborative analysis result.
2. 2. The cross-department data collaborative analysis method fusing federal learning according to claim 1, wherein the dynamically constructing a cross-department federal learning collaborative link, the cross-department federal learning collaborative link being configured to associate each department local data processing unit with a federal learning collaborative hub, generates a cross-department federal learning collaborative link configuration scheme, comprises: Collecting functional attribute information and data transmission interface information of local data processing units of each department, extracting core attribute elements representing data processing capacity in the functional attribute information of the local data processing units of each department, extracting interface attribute elements representing interface adaptation characteristics in the data transmission interface information of the local data processing units of each department, and generating core attribute and interface attribute sets of the local data processing units of each department; Acquiring a cooperative interaction requirement and a data receiving interface specification of a federal learning cooperative center, extracting interactive adaptation elements representing cooperative capacity adaptation in the cooperative interaction requirement of the federal learning cooperative center, extracting docking elements representing interface docking requirements in the data receiving interface specification of the federal learning cooperative center, and generating a federal learning cooperative center interactive adaptation and interface docking element set; performing association matching on core attributes and interface attribute sets of local data processing units of all departments and the federal learning cooperative center interactive adaptation and interface docking element sets, determining potential cooperative link nodes between the local data processing units of all departments and the federal learning cooperative center, and generating a cross-department potential cooperative link node set; carrying out link connectivity analysis on each potential cooperative link node in the cross-department potential cooperative link node set, testing whether a data transmission path between each potential cooperative link node is unobstructed or not in a simulated data transmission mode, and generating a connectivity analysis result of each potential cooperative link node; Based on the connectivity analysis result of each potential cooperative link node, screening out effective cooperative link nodes with connectivity meeting federal learning cooperative requirements, and generating a cross-department effective cooperative link node set; According to the network topology position information of the local data processing units of each department and the network topology position information of the federal learning cooperative center, the logic connection paths among the cross-department effective cooperative link nodes are planned by combining preset network performance parameters among the nodes, and the trend and node distribution of the logic connection paths are optimized to reduce data transmission delay, so that a cross-department cooperative link connection path planning scheme is generated; based on a cross-department effective cooperative link node set and a cross-department cooperative link connection path planning scheme, a link management mechanism of a cross-department federal learning cooperative link is constructed, wherein the link management mechanism comprises a link establishment flow, a data transmission rule and a link maintenance mode; combining the cross-department effective cooperative link node set, the cross-department cooperative link connection path planning scheme and the link management mechanism of the cross-department federal learning cooperative link, constructing an initial configuration framework of the cross-department federal learning cooperative link, and generating a cross-department federal learning cooperative link initial configuration framework scheme; Performing link function and performance simulation test on the cross-department federation learning cooperative link initial configuration framework scheme, and verifying the end-to-end response time delay and the effective data transmission quantity in unit time of the cross-department federation learning cooperative link under a preset scene by simulating different department data interaction scenes to generate a link function and performance simulation test result; based on the link function and performance simulation test result, adjusting node configuration parameters and path planning parameters in the cross-department federation learning cooperative link initial configuration framework scheme to generate the cross-department federation learning cooperative link configuration scheme.
3. 3. The cross-department data collaborative analysis method fusing federal learning according to claim 1, wherein the assigning federal collaborative capability to each department local data processing unit based on the cross-department federal learning collaborative link configuration scheme generates each department federal energized local data processing unit, comprising: Analyzing a cross-department federal learning cooperative link configuration scheme, extracting federal cooperative capacity requirements and link adaptation parameters related to local data processing units of all departments, and generating a federal cooperative capacity requirement and link adaptation parameter list of the local data processing units of all departments; Aiming at the existing functional modules of the local data processing units of each department, analyzing the functional gap between the existing functional modules and the federal collaborative capability requirement, identifying the types of the functional modules which need to be supplemented or upgraded, and generating a functional upgrading requirement list of the local data processing units of each department; Selecting an adaptive federal cooperative function module based on a functional ascending demand list of a local data processing unit of each department, wherein the federal cooperative function module can realize federal preprocessing of local data and cooperative data interaction, and generates an adaptive federal cooperative function module set of each department; Integrating the adaptive federal cooperative function module set of each department with the existing function module of the local data processing unit of the corresponding department, and enabling the federal cooperative function module to be capable of carrying out calling and data exchange with the existing function module according to defined interface specifications and data circulation paths by implementing the federal cooperative function module interface docking scheme and the data circulation planning scheme, so as to generate the local data processing unit of which each department preliminarily integrates federal functions; Performing parameter configuration on the local data processing units of which the federation functions are primarily integrated by each department according to federation cooperative capacity requirements and a link adaptation parameter list of the local data processing units of each department, and adjusting module operation parameters and data transmission parameters to meet the link adaptation requirements so as to generate the local data processing units after parameter configuration of each department; Performing federal collaborative function test on the local data processing units after the parameter configuration of each department, and verifying whether the local data processing units can normally execute federal data processing and collaborative data transmission or not to generate federal collaborative function test results of the local data processing units of each department; aiming at the functional defects in the federal collaborative function test result, according to the interface docking scheme of the federal collaborative function module of each department and the data flow planning scheme of the federal collaborative function module of each department, correcting interface parameters and data flow logic in the local data processing unit after parameter configuration of each department to generate a local data processing unit element after function optimization of each department; Carrying out link connection test on the local data processing units after the function optimization of each department, verifying whether the link connection between the local data processing units and the federal learning cooperative center accords with the requirements of cross-department federal learning cooperative link configuration scheme, and generating link connection test results of the local data processing units of each department; Based on the link connection test result, adjusting configuration parameters and data encapsulation format definitions related to the data transmission interface in the local data processing unit after the function optimization of each department so as to meet the requirements of cross-department federal learning collaborative link configuration scheme and generate a local data processing unit after the link adaptation adjustment of each department; And verifying the local data processing unit after the link adaptation adjustment of each department, confirming that the local data processing unit can establish connection with the federal learning cooperative center and exchange data according to the cross-department federal learning cooperative link configuration scheme, connecting the interrupt times in the test period to be lower than a preset threshold value, and conforming the processing rate of the standard test data set to the preset threshold value, and generating federal energized local data processing units of each department when all the confirmations pass.
4. 4. The method for cross-department data collaborative analysis fusing federal learning according to claim 1, wherein generating a cross-department federal collaborative data link stream through interaction of each department federally energized local data processing unit with a federal learning collaboration hub comprises: The method comprises the steps that local data to be cooperatively analyzed by a local department are collected by a federal energized local data processing unit of each department, federal preprocessing is carried out on the local data to be cooperatively analyzed by the local department, the federal preprocessing comprises data cleaning and data format standardization, and local cooperative data after preprocessing by each department are generated; The federation type local data processing unit of each department performs data encapsulation on the preprocessed local cooperative data according to the data transmission rule in the cross-department federation learning cooperative link configuration scheme to generate federation cooperative data encapsulated by each department; the federation type local data processing unit of each department transmits the encapsulated federation data to the federation learning coordination center through the established cross-department federation learning coordination link to generate federation data transmission records of each department; the federation learning cooperative center receives the encapsulated federation cooperative data transmitted by the federation energized local data processing unit of each department, performs data decapsulation on the received encapsulated federation cooperative data, extracts core cooperative data content therein, and generates a core cooperative data set received by the federation learning cooperative center; The federal learning collaboration center performs cross-department data association analysis on the core collaboration data set, identifies internal association relations among different department core collaboration data, and generates cross-department data association relation description; based on the cross-department data association relation description, the federation learning collaboration center performs data recombination on the core collaboration data sets, integrates the core collaboration data of different departments into collaboration data sets according to the association relation, and generates cross-department collaboration data set sets; the federal learning cooperative center feeds back the cross-department cooperative data group set to the corresponding federal energy-enabling local data processing unit of each department, so that the federal energy-enabling local data processing unit of each department can acquire the relevant cooperative data of other departments to generate a cross-department cooperative data feedback record; The federal energy-enabling local data processing unit of each department receives a cross-department collaborative data set fed back by the federal learning collaborative center, fuses the cross-department collaborative data set with the local collaborative data preprocessed by the department, and generates a collaborative data set fused by each department; the federal energized local data processing unit of each department transmits the fused collaborative data set to the federal learning collaborative center again to form a cross-department data collaborative interaction closed loop, and a cross-department data collaborative interaction closed loop record is generated; And integrating the data transmission path, the data content and the interaction time sequence information based on a plurality of data interaction processes between the federal energized local data processing unit and the federal learning cooperative center of each department, and generating a cross-department federal cooperative data link stream.
5. 5. The cross-department data collaborative analysis method fusing federal learning according to claim 1, wherein the constructing a cross-department data collaborative analysis model based on the cross-department federal collaborative data link flow, completing cross-department data collaborative analysis using the cross-department data collaborative analysis model, and outputting a cross-department data collaborative analysis result, comprises: Respectively extracting association features representing interaction dynamic data interaction features and data relationship from cross-department federation cooperative data link flows, and taking the data interaction features and the association features as different dimensions to jointly form a cross-department federation cooperative data feature set, wherein the data interaction features comprise data transmission frequency in terms of interaction times in unit time, data interaction delay average value in terms of time units and data transmission quantity change trend in terms of data quantity units; Determining a core architecture of a cross-department data collaborative analysis model based on a cross-department federation collaborative data feature set, wherein the core architecture is required to adapt to dynamic interaction characteristics and association relations of the cross-department data to generate a cross-department data collaborative analysis model architecture construction scheme; Selecting a model component capable of processing dynamic interaction data and complex associated data according to a cross-department data collaborative analysis model architecture construction scheme, wherein the model component comprises a dynamic data processing component and an associated data mining component, and generating a cross-department data collaborative analysis model component set; Dividing the cross-department federation cooperative data feature set into a model training feature subset and a model application feature subset, wherein the model training feature subset is used for model construction and optimization, and the model application feature subset is used for actual cooperative analysis application of the model to generate a cross-department federation cooperative data feature division result; training an initial cross-department data collaborative analysis model constructed based on a cross-department data collaborative analysis model assembly set by utilizing a model training feature subset, and performing iterative training by adjusting operation parameters of assemblies in the cross-department data collaborative analysis model assembly set to generate a trained cross-department data collaborative analysis model; inputting the model application feature subset into a trained cross-department data collaborative analysis model, testing the cross-department data collaborative analysis capability of the trained cross-department data collaborative analysis model, and generating a model test result; based on the model test result, performing fine adjustment on the operation parameters of the trained cross-department data collaborative analysis model to generate an optimized cross-department data collaborative analysis model; collecting latest data to be collaborative analysis of each department, performing federal preprocessing and packaging on the latest data to be collaborative analysis by using federal energized local data processing units of each department, and generating latest packaged collaborative data of each department; inputting the latest packaged collaborative data of each department into an optimized cross-department data collaborative analysis model, and processing input data by the optimized cross-department data collaborative analysis model through an integrated dynamic data processing component and an associated data mining component thereof to generate an initial cross-department data collaborative analysis result; and feeding back the initial cross-department data collaborative analysis result to the federal energy-forming local data processing unit and the federal learning collaborative center of each department, receiving feedback data from the federal energy-forming local data processing unit and the federal learning collaborative center of each department, and correcting the initial cross-department data collaborative analysis result based on the feedback data to generate a cross-department data collaborative analysis result.
6. 6. The federal learning fused cross-department data collaborative analysis method according to claim 2, wherein the link connectivity analysis is performed on each potential collaborative link node in a cross-department potential collaborative link node set, whether a data transmission path between each potential collaborative link node is unobstructed is tested by simulating data transmission, and a result of the link connectivity analysis of each potential collaborative link node is generated, comprising: constructing a link connectivity simulation test environment according to network parameters of an actual cross-department data transmission network, and generating a link connectivity simulation test environment configuration scheme; Setting up a simulation test platform based on a link connectivity simulation test environment configuration scheme, and re-etching all potential cooperative link nodes in a cross-department potential cooperative link node set and preset connection relations among the potential cooperative link nodes in the simulation test platform to generate a cross-department potential cooperative link simulation test platform; Generating simulation test data, wherein the format, the size and the data type of the simulation test data are consistent with those of data in an actual cross-department collaborative analysis process, and generating a cross-department link connectivity simulation test data set; Transmitting a cross-department link connectivity simulation test data set from a starting potential cooperative link node in a simulation test platform to a target potential cooperative link node according to a preset transmission time sequence, and recording transmission state information in a data transmission process; Extracting key indexes in transmission state information in a data transmission process, including whether data transmission is successful, data loss condition and transmission delay time length in the transmission process, and generating a data transmission key index set of each potential cooperative link node; Analyzing the data transmission key index set of each potential cooperative link node, judging whether the data transmission between each potential cooperative link node can meet the basic requirements of cross-department federal learning cooperation, and generating transmission suitability judgment results of each potential cooperative link node; aiming at potential cooperative link nodes with unsuccessful data transmission or data loss and unsatisfactory transmission delay time, readjusting simulation test parameters to carry out repeated tests for a plurality of times, and generating transmission state information after repeated tests for a plurality of times; Integrating transmission state information after single test and repeated test for multiple times, comprehensively judging connectivity stability of each potential cooperative link node, and generating a connectivity stability evaluation result of each potential cooperative link node; based on the data transmission key index set, the transmission suitability judgment result and the connectivity stability evaluation result of each potential cooperative link node, summarizing to form a connectivity analysis result of each potential cooperative link node; Classifying and sorting the connectivity analysis results of the potential cooperative link nodes, clearly marking the potential cooperative link nodes with the connectivity meeting the requirements and not meeting the requirements, and generating a structured connectivity analysis result report of each potential cooperative link node.
7. 7. The method for collaborative analysis of cross-department data fusing federal learning according to claim 3, wherein the integrating the set of adaptive federal collaborative function modules of each department with existing functional modules of local data processing units of corresponding departments enables the federal collaborative function modules to call and exchange data with the existing functional modules according to defined interface specifications and data flow paths by implementing the federal collaborative function module interface docking scheme and the data flow planning scheme, and generates the local data processing units of the initial integrated federal functions of each department, comprising: acquiring interface definition documents and data flow logic descriptions of existing functional modules of local data processing units of all departments, extracting input interface parameters, output interface parameters and interface calling modes of the existing functional modules, and simultaneously extracting data flow paths and data processing sequences among the existing functional modules to generate an interface and flow logic set of the existing functional modules of the local data processing units of all departments; Analyzing the interface specification and the data processing logic of each federal cooperative function module in each department adaptive federal cooperative function module set, extracting the input and output interface parameters, the interface adaptive requirements and the data processing flow of the federal cooperative function module, and generating each department adaptive federal cooperative function module interface and processing logic set; Performing interface matching analysis on the interface and processing logic set of the adaptive federation cooperative function module of each department and the interface and circulation logic set of the existing function module of the local data processing unit of the corresponding department, determining an interface docking mode between the federation cooperative function module and the existing function module, and generating an interface docking scheme of the federation cooperative function module of each department; Based on the interface docking scheme of the federation cooperative function module of each department, adjusting interface parameters of the federation cooperative function module to generate a federation cooperative function module with the interfaces adjusted by each department; Planning a data flow path between the federal cooperative function module and the existing function module, and defining a specific flow of data flowing from the existing function module to the federal cooperative function module and from the federal cooperative function module to the existing function module to generate a data flow planning scheme of the federal cooperative function module of each department; according to the data flow planning scheme of the federal cooperative function module of each department, the data flow logic of the existing function module is adjusted, a data flow channel between the federal cooperative function module and the existing function module is newly added, and a local data processing unit with the adjusted data flow logic of each department is generated; Embedding the federation cooperative function module with the adjusted interface into the local data processing unit with the adjusted data flow logic to realize the physical integration of the federation cooperative function module and the existing function module and generate the local data processing unit after the physical integration of each department; Carrying out data flow test on the local data processing unit after physical integration of each department, verifying whether the data flow between the federal cooperative function module and the existing function module is smooth or not through input test data, and generating a data flow test result of each department; aiming at the problems of flow blocking or incompatibility of data formats in the data flow test result, the interface parameters and the data flow logic are optimized and adjusted again to generate an integrated unit with optimized interfaces and flow logic of each department; and carrying out functional integrity test on the integrated units with optimized interfaces and circulation logic of each department, verifying whether the integrated units can completely realize original functions and newly added federal cooperative functions, and generating a local data processing unit with preliminarily integrated federal functions of each department.
8. 8. The cross-department data collaborative analysis method fusing federal learning according to claim 4, wherein the federal learning collaboration center performs cross-department data association analysis on a core collaboration data set, identifies internal association relationships between different department core collaboration data, and generates cross-department data association relationship descriptions, comprising: Extracting data characteristics of each piece of core cooperative data in the core cooperative data set, and extracting attribute characteristics, content characteristics and source characteristics of each piece of core cooperative data, wherein the attribute characteristics represent types and purposes of the data, the content characteristics represent core information of the data, and the source characteristics represent departments to which the data belong to, so as to generate the core cooperative data characteristic set; Classifying the core collaborative data sets according to department sources based on source features in the core collaborative data feature sets to generate core collaborative data subsets of each department; Carrying out internal data characteristic association analysis on the core cooperative data subsets of each department, identifying association relations among different core cooperative data in the same department, and generating an internal data association relation set of each department; selecting a core cooperative data subset of one department as a reference data subset, comparing the characteristics of the core cooperative data subset of other departments with the reference data subset, searching common characteristics and similar characteristics among the core cooperative data of different departments, and generating a cross-department data characteristic comparison result; based on the cross-department data feature comparison result, extracting common features and similar features among different department core cooperative data, and pairing the core cooperative data with the common features or the similar features to generate a cross-department core cooperative data association pair set; Performing association strength analysis on each association data pair in the cross-department core collaborative data association pair set, determining association tightness degree of the association data pairs by analyzing the matching degree of the number of common features and similar features among the association data pairs, and generating a cross-department data association strength evaluation result; Classifying cross-department core collaborative data association pairs according to the cross-department data association strength evaluation result to generate a cross-department data association classification result; Combining the internal data association relation set of each department and the cross-department data association classification result to construct a cross-department data association network, wherein the network node is core cooperative data, and the network side is the association relation between the data to generate the cross-department data association network; Carrying out structural analysis on a cross-department data association network, and identifying core association nodes and key association paths in the network, wherein the core association nodes are data with association relation with a plurality of other data, and the key association paths are main paths for connecting core data of different departments to generate a cross-department data association network structural analysis result; Based on the cross-department data association classification result, the cross-department data association network and the cross-department data association network structure analysis result, the association relationship type, the association strength and the association path between different department core cooperative data are described in detail, and the cross-department data association relationship description is generated.
9. 9. The federal learning fused cross-department data collaborative analysis method according to claim 5, wherein training an initial model constructed based on a set of cross-department data collaborative analysis model components using a subset of model training features, generating a trained cross-department data collaborative analysis model by adjusting fitting capabilities of an operational parameter optimization model of the model components to cross-department data features, comprising: Dividing the model training feature subset into a model training subset and a model verification subset according to a preset proportion, wherein the model training subset is used for adjusting and training cross-department data collaborative analysis model parameters, and the model verification subset is used for verifying cross-department data collaborative analysis model performance in the training process to generate a model training and verification data dividing result; Inputting the model training subset into an initial cross-department data collaborative analysis model constructed based on a cross-department data collaborative analysis model assembly set, capturing dynamic data characteristics of the model training subset by a dynamic data processing assembly in the initial cross-department data collaborative analysis model, extracting associated characteristics of the model training subset by an associated data mining assembly, and generating a training output result of the initial cross-department data collaborative analysis model; Extracting feature fitting information in a training output result of the initial cross-department data collaborative analysis model, analyzing differences between the training output result of the initial cross-department data collaborative analysis model and actual features of a model training subset, and generating a model training fitting difference analysis result; Based on the model training fit difference analysis result, determining a cross-department data collaborative analysis model component and a corresponding adjustment direction which need to be adjusted, adjusting dynamic characteristic capturing parameters of the cross-department data collaborative analysis model component aiming at a dynamic data processing component, adjusting associated characteristic extraction parameters of the cross-department data collaborative analysis model component aiming at an associated data mining component, and generating a cross-department data collaborative analysis model parameter adjustment scheme; adjusting related component parameters of the initial cross-department data collaborative analysis model according to a cross-department data collaborative analysis model parameter adjustment scheme, and generating a cross-department data collaborative analysis model with adjusted parameters; inputting the model verification subset into the cross-department data collaborative analysis model after parameter adjustment, obtaining a verification output result of the cross-department data collaborative analysis model after parameter adjustment, analyzing a fitting difference between the verification output result and actual characteristics of the model verification subset, and generating a model verification fitting difference analysis result; Judging whether the model verification fit difference analysis result accords with a preset cross-department data collaborative analysis model training termination condition, if not, re-optimizing a cross-department data collaborative analysis model parameter adjustment scheme based on the verification fit difference analysis result, and repeatedly executing parameter adjustment and cross-department data collaborative analysis model verification flow; in the verification process of multiple parameter adjustment and cross-department data collaborative analysis model, recording fitting differences between cross-department data collaborative analysis model parameters after each adjustment and corresponding cross-department data collaborative analysis models, and generating a cross-department data collaborative analysis model training process record; Stopping the parameter adjustment flow when the model verification fitting difference analysis result accords with a preset cross-department data collaborative analysis model training termination condition, and selecting a cross-department data collaborative analysis model parameter combination with the minimum fitting difference as a final cross-department data collaborative analysis model parameter to generate a cross-department data collaborative analysis model final parameter set; And configuring a final parameter set of the cross-department data collaborative analysis model into the initial cross-department data collaborative analysis model to generate a trained cross-department data collaborative analysis model.
10. 10. The cross-department data collaborative analysis system for fusion federal learning is characterized by comprising a processor and a memory, wherein the memory is connected with the processor, the memory is used for storing programs, instructions or codes, and the processor is used for running the programs, instructions or codes in the memory so as to realize the cross-department data collaborative analysis method for fusion federal learning according to any one of claims 1-9.

Description

Cross-department data collaborative analysis method and system integrating federal learning Technical Field The invention relates to the technical field of federal learning, in particular to a cross-department data collaborative analysis method and system for federal learning. Background In the current digital age, various departments accumulate a large amount of valuable data in the process of business development, and the data contains rich information, so that the method has important significance for improving decision science, optimizing business processes and the like. However, due to the fact that the data between departments often have many problems such as data privacy protection, data security compliance and data format difference, the data of each department are difficult to directly share and integrate, and each data island is formed. The existing data collaborative analysis method has obvious defects in solving the cross-department data collaborative problem. On the one hand, some centralized data processing methods need to concentrate the data of each department to a central server for processing, which not only faces security risks in the data transmission process, but also may violate the data privacy protection regulations, because the data of each department may contain sensitive information, such as personal identity information, business confidentiality, and the like. On the other hand, although the traditional distributed data processing method considers the distributed storage of data to a certain extent, when cross-department collaborative analysis is performed, an effective collaborative mechanism and a unified model construction method are lacked, effective fusion and deep analysis of data of each department are difficult to realize, and the comprehensive value of the cross-department data cannot be fully exerted, so that the application effect of the data in cross-department business decision and optimization is limited. Disclosure of Invention In view of the above-mentioned problems, in combination with the first aspect of the present invention, an embodiment of the present invention provides a cross-department data collaborative analysis method for fusion federal learning, the method including: Dynamically constructing a cross-department federation learning cooperative link, wherein the cross-department federation learning cooperative link is used for associating local data processing units of all departments with federation learning cooperative centers to generate a cross-department federation learning cooperative link configuration scheme; The federal collaborative capability is given to the local data processing units of each department based on the cross-department federal learning collaborative link configuration scheme, and federal energized local data processing units of each department are generated; Generating cross-department federation cooperative data link flows through interaction between federation energized local data processing units of all departments and federation learning cooperative centers; and constructing a cross-department data collaborative analysis model based on the cross-department federation collaborative data link flow, completing cross-department data collaborative analysis by using the cross-department data collaborative analysis model, and outputting a cross-department data collaborative analysis result. In yet another aspect, an embodiment of the present invention further provides a cross-department data collaborative analysis system that fuses federal learning, including a processor, a machine-readable storage medium, where the machine-readable storage medium is connected to the processor, and the machine-readable storage medium is used to store a program, an instruction, or a code, and the processor is used to run the program, the instruction, or the code in the machine-readable storage medium, so as to implement the method described above. Based on the above aspects, the embodiment of the invention establishes an efficient and safe association channel for each department local data processing unit and the federal learning coordination center by dynamically constructing the cross-department federal learning coordination link and generating a corresponding configuration scheme, thereby solving the problem of the basic architecture of cross-department data coordination. Based on the configuration scheme, federal collaborative capability is given to the local data processing units of each department, and federal energized local data processing units are generated, so that each department can participate in federal learning on the premise of not revealing original data, and the data privacy and safety are effectively ensured. The cross-department federation cooperative data link flow is generated through interaction of the federation energized local data processing units of each department and the federation learning cooperative center, ordered fl