Multi-field business collaborative management system and method based on block chain

Abstract

The invention discloses a multi-field business collaborative management system and method based on a block chain, and relates to the technical field of business collaborative management. Standardization rules are introduced, the field name, the data type and the unique identifier of data are processed in a unified manner, and by labeling sources, timestamps and operation records, the business collaborative management efficiency is improved; an anomaly detection algorithm based on statistical distribution, machine learning and rule verification is adopted, consistency verification and repeated data cleaning are combined, key features are extracted through principal component analysis, sparse coding and a deep learning embedding model, semantic alignment and feature fusion of texts, numerical values, time sequences and image data are achieved through multi-modal modeling, and the feature fusion of the text, numerical values, time sequences and image data is achieved. Cross-modal unified data representation is generated; a knowledge graph technology and a block chain technology are combined, a semantic relationship between fields is constructed through semantic alignment, an alliance chain is used for recording data circulation and operation, dynamic access authority management is realized in combination with an intelligent contract, and transparency and non-tampering performance of data are guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of business collaboration management, and in particular to a multi-domain business collaboration management system and method based on blockchain. Background Art

[0002] In the multi-domain business management of modern enterprises, collaboration has become a key link to promote the efficient operation of organizations. Whether it is the close connection between the upstream and downstream of the supply chain or the seamless linkage of the R & D, production, and sales links, enterprises are committed to achieving the rational allocation of resources and the rapid response to market demands through cross-domain cooperation.

[0003] The core of multi-domain business collaboration lies in the sharing and integration of data. However, in reality, the sources of data are diverse, the structures are complex, and the distribution is extensive; the data standards and technical protocols between different departments and business links are often not unified, resulting in obstacles in the integration process; to bridge these gaps, enterprises need to invest a large amount of human and technical resources and rely on manual processing or intermediate system docking to achieve the initial docking of data; however, this approach is not effective in scenarios with high real-time requirements; at the same time, data integration not only needs to deal with the differences between structured and unstructured data but also needs to make trade-offs in terms of quality and credibility; due to the opacity of the data collection process and the possible tampering or loss in intermediate links, all parties participating in the collaboration lack trust in the authenticity and reliability of the shared data, further exacerbating the complexity of collaboration; this trust crisis forces many enterprises to set up redundant links in data processing to avoid risks. However, this undoubtedly increases the additional time cost and operation burden.

[0004] To address the above problems, some traditional methods establish a unified data warehouse or build a middleware architecture for centralized data management and transfer, which alleviates the data island problem to a certain extent. However, once a system failure or external attack occurs, the entire collaboration network will face the risk of paralysis; in addition, in the face of the increasing data scale and dynamic business requirements, the lack of flexibility and scalability of the centralized mode becomes increasingly prominent; enterprises are difficult to efficiently handle these increasingly complex collaboration requirements, resulting in a decline in the efficiency of business collaboration and an impact on competitiveness; in this context, how to break through the limitations of the traditional collaboration management mode and solve the problems of low data integration efficiency, insufficient credibility, and limited scalability has become a core challenge to be solved urgently. Summary of the Invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] The present invention provides a multi-domain business collaboration management system and method based on blockchain to solve the problems of low data integration efficiency, insufficient credibility, and limited scalability existing in traditional collaboration management solutions.

[0007] To solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, an embodiment of the present invention provides a multi-domain business collaboration management system based on blockchain, which includes

[0009] A data collection module, responsible for collecting data from heterogeneous data sources in multiple domains, annotating the data source, timestamp, and operation records, and defining standardization rules, including field names, data types, and unique identifiers;

[0010] A data cleaning module, which preprocesses the data collected by the data collection module, uses anomaly detection algorithms to remove invalid data, uses quality verification techniques for consistency verification, integrity check, and duplicate data cleaning, and converts JSON, XML, and CSV heterogeneous data formats into a standardized format through formatting tools;

[0011] A data fusion module, which extracts feature data from the preprocessed data, correlates cross-domain data through modeling techniques, and semantically aligns text, numerical, time series, and image data using multi-modal modeling techniques to generate fusion data;

[0012] A trusted guarantee module, which constructs a knowledge graph based on the fusion data, performs semantic alignment, and introduces blockchain technology. The blockchain adopts a consortium chain structure and combines smart contracts to dynamically manage data access permissions, while recording the flow and operation of data;

[0013] A collaborative management platform, used to receive the knowledge graph and manage multi-domain businesses.

[0014] As a preferred solution of the multi-domain business collaboration management system based on blockchain described in the present invention, wherein: in the collaborative management platform, the business process is optimized in real time by combining intelligent algorithms, and the graph data is displayed through a visual dashboard and analysis views.

[0015] In a second aspect, the present invention provides a multi-domain business collaboration management method based on blockchain, including

[0016] Step S1, perform cross-domain data collection, and annotate the data source, timestamp, and operation records during the collection process;

[0017] During the data collection process, standardize the field names, data types, and unique identifiers of the data;

[0018] Step S2, preprocess the data collected in step S1;

[0019] Step S3: Through feature extraction and multi-modal modeling techniques, extract features and associate features with the data preprocessed in Step S2 to generate fused data;

[0020] Step S4: Based on the fused data, use knowledge graph technology to generate semantic relationships between domains and construct a knowledge graph, and at the same time introduce blockchain technology to record data flow and operations;

[0021] The knowledge graph realizes cross-domain relationship modeling through semantic alignment;

[0022] The blockchain technology adopts a consortium chain structure and combines smart contracts to dynamically manage data access permissions;

[0023] Step S5: Embed the generated knowledge graph into the collaborative management platform and perform real-time analysis and dynamic optimization in combination with intelligent algorithms;

[0024] The intelligent algorithms include a dynamic optimization model, a prediction algorithm, and a resource scheduling algorithm;

[0025] The collaborative management platform real-time displays business status and data through a visual dashboard and an analysis view.

[0026] As a preferred solution of the multi-domain business collaborative management method based on blockchain described in the present invention, wherein: the step of standardizing the field name, data type, and unique identifier of the data is,

[0027] Define the original collected data set as D: D = {d 1 , d 2 ,..., d n},

[0028] wherein, D represents the original collected data set, d i represents the i-th data entry, and the composition form of each data entry d i is: d i = (f i , t i , u i ), wherein, f i represents the field name, t i represents the data type, and u i represents the unique identifier,

[0029] Define the standardization rule set as R:

[0030] wherein, R represents the standardization rule set, represents the standardized field name, represents the standardized data type, represents the standardized unique identifier;

[0031] Apply the rule set R to the data set D to generate the standardized data set D s :

[0032]

[0033] where D s represents the standardized data set, are respectively the field name, data type, and unique identifier after standardization,

[0034] Perform conflict detection, and the detection function is:

[0035]

[0036] where Q(D s ) represents the result of the conflict detection function, represents the pair of field names where a conflict occurs, represents the type of the conflicting field, represents the unique identifier of the conflicting field. If Update the standardization rule R according to the conflict correction rule and regenerate D s , until

[0037] As a preferred solution of the multi-domain business collaboration management method based on blockchain described in the present invention, where: the preprocessing includes: using an anomaly detection algorithm and a quality verification technology to remove invalid data, and converting heterogeneous data into a standardized format through a formatting tool;

[0038] The anomaly detection algorithm includes an analysis algorithm based on statistical distribution, an anomaly point detection model based on machine learning, and a boundary condition verification driven by rules;

[0039] The quality verification technology includes consistency verification, integrity check, and duplicate data cleaning;

[0040] The formatting tool supports automatic mapping and standardization processing of JSON, XML, and CSV data formats.

[0041] As a preferred solution of the multi-domain business collaboration management method based on blockchain described in the present invention, where: the step of using an anomaly detection algorithm and a quality verification technology to remove invalid data is,

[0042] Assume the data set is Z: Z = {Z 1 , Z 2 ,..., Z k},

[0043] where Z represents the data set to be processed, Zi Denote the i-th data point in the dataset,

[0044] Calculate the probability density of the distribution. The probability density function is:

[0045]

[0046] where, P z (z) represents the distribution probability density of the data point z, μ z represents the mean of the dataset, σ z represents the standard deviation of the dataset, and exp represents the exponential function.

[0047] Determine the set of outlier points. The determination formula is:

[0048] Z a = {z r | P z (z r ) < ∈ z}},

[0049] where, Z a represents the set of outlier points, z r represents the r-th data point, and ∈ z represents the threshold for outlier determination;

[0050] Use the Isolation Forest model F z to train the dataset Z and calculate the outlier score for each data point. The calculation formula is:

[0051]

[0052] where, S z (z t ) represents the outlier score of the data point z t , h z (z t ) represents the average splitting depth of the data point in the Isolation Forest, and max(h z (Z)) represents the maximum value of all splitting depths in the dataset.

[0053] Determine the outlier data points. The determination formula is:

[0054] Z f = {Z t | S z (z t ) > T z}},

[0055] where, Z f represents the set of data determined to be outliers, and τ z represents the threshold of the outlier score.

[0056] Perform consistency verification and define the set of field values as V:

[0057] V = {v 1 , v 2 ,..., v q},

[0058] where V represents the set of values of all data fields, and v i represents the value of the i-th field,

[0059] Define the consistency rule function:

[0060] V invalid = {v s | C v (v s ) = False},

[0061] where C v (v s ) represents the consistency rule function, V invalid represents the set of field values that do not satisfy the consistency rule, and False represents the determination result that the field value does not satisfy the consistency rule,

[0062] Define the data point similarity function:

[0063]

[0064] where Sim z (z u , z v ) represents the similarity between data points z u and z v , |z u - z v | represents the Euclidean distance between data points, and max(|z u - z v |) represents the maximum distance in the dataset,

[0065] If Sim z (z u , z v ) > λ z , then retain z u and delete z v .

[0066] As a preferred solution of the multi-domain business collaboration management method based on blockchain described in the present invention, wherein: the feature extraction technology includes principal component analysis PCA based on dimensionality reduction, sparse coding, and deep learning embedding model;

[0067] The multi-modal modeling uses a data semantic alignment model to associate text, numerical, time series, and image data to form fused data.

[0068] As a preferred solution of the method for multi-domain business collaborative management based on blockchain according to the present invention, wherein: the step of performing feature extraction and feature association on the data preprocessed in step S2 is

[0069] Perform feature extraction, use principal component analysis PCA for dimensionality reduction, and define the data matrix as X:

[0070] X = [x 1 , x 2 , …, x k T ,

[0071] wherein, X is an n×k data matrix containing k samples, x i represents the data vector of the i-th sample, and T is the transpose operation.

[0072] Calculate the covariance matrix, and the calculation formula is:

[0073]

[0074] wherein,

[0075] ∑ X represents the covariance matrix of the data, μ X is the mean vector of the data.

[0076] Perform eigen-decomposition on the covariance matrix to obtain the principal components, and the decomposition formula is:

[0077] ∑ X v j = λ j v j ,

[0078] wherein, v j is the j-th eigenvector, and λ j is the corresponding eigenvalue.

[0079] Extract the first m principal components, and the data after dimensionality reduction is X′:

[0080] X′ = X·V m ,

[0081] wherein, X′ is the data matrix after dimensionality reduction, and V m is the projection matrix composed of the first m eigenvectors;

[0082] Perform feature association, and define the multi-modal data as M:

[0083] M = {T, N, S, I},​

[0084] Among them, M represents the multimodal data set, T is the text data, N is the numerical data, S is the time series data, and I is the image data.

[0085] Use the semantic alignment function to associate multimodal data, and the association formula is:

[0086] R M = φ(T, N, S, I),

[0087] Among them, R M is the semantic alignment result, and φ represents the multimodal association function.

[0088] Fuse the aligned data features, and the fusion method is:

[0089] F M = Concat(X T , X N , X S , X I ),

[0090] Among them, F M is the fused feature data, and X T , X N , X S , X I are the feature representations of text, numerical value, time series, and image respectively, and Concat represents the feature concatenation operation.

[0091] As a preferred solution of the multi-domain business collaboration management method based on blockchain described in the present invention, among them: the steps of using the knowledge graph technology to generate the semantic relationship between domains and constructing the knowledge graph, and at the same time introducing the blockchain technology to record data flow and operations are as follows.

[0092] Construct a knowledge graph, define nodes and relationships, and define the node set of the knowledge graph as V:

[0093] V = {v 1 , v 2 ,..., v m},

[0094] Among them, V is the node set, and v i represents the i-th node.

[0095] Define the relationship set as E:

[0096] E = {e 1 , e 2 ,..., e n},

[0097] Among them, E is the relationship set, and ei Denote the i-th relationship,

[0098] Use the semantic alignment function to calculate the semantic weight between nodes, and the calculation formula is:

[0099] R ij = ψ(v i , v j ),

[0100] where, R ij represents the semantic relationship weight between node v i and node v j , ψ represents the semantic alignment function,

[0101] Generate the knowledge graph G:

[0102] G = (V, E, R),

[0103] where, G represents the knowledge graph, and R is the set of semantic relationship weights;

[0104] Introduce blockchain technology to generate the blockchain hash value for the knowledge graph data G, and the hash formula is:

[0105] H G = Hash(G),

[0106] where, H G represents the blockchain hash value of the knowledge graph, and Hash represents the hash function,

[0107] Use the smart contract to dynamically manage the data access permission, and the management formula is:

[0108] P c = Contract(G, A),

[0109] where, P c represents the permission control policy, Contract is the smart contract, and A is the set of access permissions.

[0110] As a preferred solution of the multi-domain business collaboration management method based on blockchain described in the present invention, wherein: the step of combining the intelligent algorithm for real-time analysis and dynamic optimization is,

[0111] Define the data set X and the prediction model P x :

[0112]

[0113] where, represents the prediction result, P x represents the prediction model,

[0114] Define the objective function as J(x):

[0115] J(x) = g(x) - h(x),

[0116] where J(x) is the objective function, g(x) represents the revenue function, and h(x) represents the cost function,

[0117] The optimal solution is obtained by an optimization algorithm, and the formula is:

[0118] x * = arg max J(x),

[0119] where x * represents the optimal solution of the objective function.

[0120] The beneficial effects of the present invention are as follows: In the present invention, a standardization rule is introduced to uniformly process the field names, data types, and unique identifiers of data, and through the annotation of sources, timestamps, and operation records, the standardization and traceability of data are realized, solving the problem of inconsistent data standards in traditional solutions; secondly, in the data cleaning module, an anomaly detection algorithm based on statistical distribution, machine learning, and rule verification is adopted, combined with consistency verification and duplicate data cleaning, significantly improving the quality and reliability of data. In the data fusion module, key features are extracted through principal component analysis, sparse coding, and deep learning embedding models, and semantic alignment and feature fusion of text, numerical, time series, and image data are realized by multi-modal modeling, generating a cross-modal unified data representation, effectively solving the problem of the complexity of cross-domain data association and fusion; the trusted guarantee module combines knowledge graph technology and blockchain technology, constructs semantic relationships between domains through semantic alignment, records data flow and operations with a consortium chain, and realizes dynamic access permission management in combination with smart contracts, ensuring the transparency and immutability of data, and alleviating the problem of insufficient data credibility in the traditional mode; the collaborative management platform analyzes and optimizes the business process in real time through intelligent algorithms, and intuitively presents the optimization results through a visual dashboard, providing real-time support for dynamic decision-making.

[0121] In summary, the present invention effectively overcomes the problems of low data integration efficiency, insufficient data credibility, vulnerability of the centralized management mode, and poor scalability in traditional collaborative management, and has significant advantages in terms of system reliability, flexibility, and business response speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0122] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0123] Figure 1 It is a schematic diagram of the framework of the multi - domain business collaboration management system based on blockchain of the present invention.

[0124] Figure 2 It is a schematic diagram of the process of the multi - domain business collaboration management method based on blockchain of the present invention. Detailed implementation manners

[0125] To make the above - mentioned objects, features and advantages of the present invention more obvious and understandable, the following will describe the detailed implementation manners of the present invention with reference to the accompanying drawings of the specification.

[0126] In the following description, many specific details are set forth to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Those skilled in the art can make similar promotions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0127] Secondly, the so - called "one embodiment" or "embodiment" herein refers to a specific feature, structure or characteristic that can be included in at least one implementation manner of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments.

[0128] Embodiment 1, referring to Figure 1 and Figure 2 , this embodiment provides a multi - domain business collaboration management system based on blockchain, including:

[0129] A data acquisition module, which is responsible for collecting data from heterogeneous data sources in multiple domains, annotating the data source, timestamp and operation record, and defining standardization rules, including field names, data types and unique identifiers;

[0130] A data cleaning module, which pre - processes the data collected by the data acquisition module, uses anomaly detection algorithms to remove invalid data, uses quality verification techniques for consistency verification, integrity check and duplicate data cleaning, and converts JSON, XML and CSV heterogeneous data formats into a standardized format through formatting tools;

[0131] A data fusion module, which extracts feature data from the pre - processed data, correlates cross - domain data through modeling techniques, and uses multi - modal modeling techniques to perform semantic alignment on text, numerical, time - series and image data to generate fusion data;

[0132] A trusted guarantee module, which constructs a knowledge graph based on the fusion data, performs semantic alignment, and introduces blockchain technology. The blockchain adopts a consortium chain structure and combines smart contracts to dynamically manage data access permissions, and at the same time records the flow and operation of data;

[0133] A collaborative management platform for receiving a knowledge graph and managing multi-domain operations;

[0134] In the collaborative management platform, intelligent algorithms are combined to optimize business processes in real time, and the graph data is presented through a visual dashboard and analysis views.

[0135] This embodiment also provides a multi-domain business collaboration management method based on blockchain, including:

[0136] Step S1: Conduct cross-domain data collection, and mark the data source, timestamp, and operation records during the collection process;

[0137] During the data collection process, standardize the field names, data types, and unique identifiers of the data;

[0138] The steps for standardizing the field names, data types, and unique identifiers of the data are as follows:

[0139] Define the original collected data set as D: D = {d 1 , d 2 ,..., d n},

[0140] where D represents the original collected data set, and d i represents the i-th data entry. The composition form of each data entry d i is: d i = (f i , t i , u i ), where f i represents the field name, t i represents the data type, and u i represents the unique identifier.

[0141] Define the standardization rule set as R:

[0142] where R represents the standardization rule set, represents the standardized field name, represents the standardized data type, represents the standardized unique identifier;

[0143] Apply the rule set R to the data set D to generate the standardized data set D s :

[0144]

[0145] where D s represents the standardized data set. They are the standardized field names, data types, and unique identifiers respectively,

[0146] Perform conflict detection, and the detection function is:

[0147]

[0148] where Q(D s ) represents the result of the conflict detection function, represents the pair of field names where a conflict occurs, represents the type of the conflicting field, represents the unique identifier of the conflicting field. If Update the standardization rule R according to the conflict correction rule and regenerate D s , until

[0149] Specifically, through the standardization rule and conflict detection, gradually standardize the field names, data types, and unique identifiers of the collected data, and unify the data structure.

[0150] Step S2, preprocess the data collected in step S1;

[0151] The preprocessing includes: using anomaly detection algorithms and quality verification technologies to remove invalid data, and converting heterogeneous data into a standardized format through a formatting tool;

[0152] The anomaly detection algorithms include analysis algorithms based on statistical distributions, outlier detection models based on machine learning, and boundary condition verification driven by rules;

[0153] The quality verification technologies include consistency verification, integrity check, and duplicate data cleaning;

[0154] The formatting tool supports automatic mapping and standardization processing of JSON, XML, and CSV data formats;

[0155] The steps of using anomaly detection algorithms and quality verification technologies to remove invalid data are,

[0156] Assume the data set is Z: Z = {z 1 , z 2 ,..., z k},

[0157] where Z represents the data set to be processed, and z i represents the i-th data point in the data set,

[0158] Calculate the probability density of the distribution, and the probability density function is:

[0159]

[0160] Among them, P z (z) represents the distribution probability density of the data point z, and μ z represents the mean of the data set, and σ z represents the standard deviation of the data set. exp represents the exponential function,

[0161] Determine the set of outlier points. The determination formula is:

[0162] Z a ={z r |P z (z r )<∈ z}},

[0163] Among them, Z a represents the set of outlier points, z r represents the r-th data point, and ∈ z represents the threshold for outlier determination;

[0164] Use the Isolation Forest model F z to train the data set Z and calculate the outlier score of each data point. The calculation formula is:

[0165]

[0166] Among them, S z (z t ) represents the outlier score of the data point z t , h z (z t ) represents the average splitting depth of the data point in the Isolation Forest, and max(h z (Z)) represents the maximum value of all splitting depths in the data set,

[0167] Determine the outlier data points. The determination formula is:

[0168] Z f ={z t |S z (z t )>T z}},

[0169] Among them, Z f represents the data set determined to be an outlier, and τ z represents the threshold of the outlier score,

[0170] Perform consistency verification. Define the set of field values as V:

[0171] V={v 1 , v 2 ,..., v q},

[0172] Among them, V represents the set of values of all data fields, and v i represents the value of the i-th field,

[0173] Define the consistency rule function:

[0174] V invalid ={v s |C v (v s ) = False},

[0175] where C v (v s ) represents the consistency rule function, V invalid represents the set of field values that do not satisfy the consistency rule, and False represents the determination result that the field value does not satisfy the consistency rule,

[0176] Define the data point similarity function:

[0177]

[0178] Among them, Sim z (z u , z v ) represents the similarity between data points z u and z v , |z u - z v | represents the Euclidean distance between data points, and max(|z v - z v |) represents the maximum distance in the dataset,

[0179] If Sim z (z u , z v ) > λ z , then retain z u and delete z v ;

[0180] Specifically, adopt the anomaly detection technology based on the distribution model and the isolation forest to efficiently identify data anomaly points, and combine the consistency verification and duplicate data cleaning technology to ensure the integrity and consistency of the data.

[0181] Step S3, through the feature extraction and multi-modal modeling technology, perform feature extraction and feature association on the data preprocessed in step S2 to generate fusion data;

[0182] The feature extraction technology includes principal component analysis PCA based on dimensionality reduction, sparse coding, and deep learning embedding models;

[0183] Multimodal modeling adopts a data semantic alignment model to associate text, numerical, time series, and image data to form fused data;

[0184] The steps of performing feature extraction and feature association on the data preprocessed in step S2 are as follows:

[0185] For feature extraction, principal component analysis (PCA) is used for dimensionality reduction. Define the data matrix as X:

[0186] X = [x 1 , x 2 , …, x k T ,

[0187] where X is an n×k data matrix containing k samples, and x i represents the data vector of the i-th sample, and T is the transpose operation.

[0188] Calculate the covariance matrix. The calculation formula is:

[0189]

[0190] where

[0191] ∑x represents the covariance matrix of the data, and μ X is the mean vector of the data.

[0192] Perform eigen decomposition on the covariance matrix to obtain the principal components. The decomposition formula is:

[0193] ∑ X v j = λ j v j ,

[0194] where v j is the j-th eigenvector, and λ j is the corresponding eigenvalue.

[0195] Extract the first m principal components. The data after dimensionality reduction is X′:

[0196] X′ = X·V m ,

[0197] where X′ is the data matrix after dimensionality reduction, and V m is the projection matrix composed of the first m eigenvectors.

[0198] For feature association, define the multimodal data as M:

[0199] M = {T, N, S, I},

[0200] ​Among them, M represents the multimodal data set, T is the text data, N is the numerical data, S is the time series data, and I is the image data.

[0201] The multimodal data is associated using a semantic alignment function, and the association formula is:

[0202] R M = φ(T, N, S, I),

[0203] Among them, R M is the semantic alignment result, and φ represents the multimodal association function.

[0204] The aligned data features are fused, and the fusion method is:

[0205] F M = Concat(X T , X N , X S , X I ),

[0206] Among them, F M is the fused feature data, and X T , X N , X S , X I are the feature representations of text, numerical, time series, and image respectively, and Concat represents the feature concatenation operation.

[0207] Specifically, dimensionality reduction is performed through principal component analysis to extract efficient features and reduce data redundancy; and multimodal modeling is carried out through semantic alignment and feature fusion technologies to associate text, numerical, time series, and image data, and generate cross-modal unified features.

[0208] Step S4: Based on the fused data, use knowledge graph technology to generate semantic relationships between domains and construct a knowledge graph, and at the same time introduce blockchain technology to record data flow and operations;

[0209] The knowledge graph realizes the modeling of cross-domain relationships through semantic alignment;

[0210] The blockchain technology adopts a consortium chain structure and combines smart contracts to dynamically manage data access permissions;

[0211] The steps of using knowledge graph technology to generate semantic relationships between domains and construct a knowledge graph, and at the same time introducing blockchain technology to record data flow and operations are as follows:

[0212] Construct a knowledge graph, define nodes and relationships, and define the node set of the knowledge graph as V:

[0213] V = {v 1 , v 2 ,..., vm},

[0214] where V is the set of nodes, and v i represents the i-th node,

[0215] define the set of relationships as E:

[0216] E = {e 1 , e 2 ,..., e n},

[0217] where E is the set of relationships, and e i represents the i-th relationship,

[0218] calculate the semantic weight between nodes using the semantic alignment function, and the calculation formula is:

[0219] R ij = ψ(v i , v j ),

[0220] where R ij represents the semantic relationship weight between node v i and node v j , and ψ represents the semantic alignment function,

[0221] generate the knowledge graph G:

[0222] G = (V, E, R),

[0223] where G represents the knowledge graph, and R is the set of semantic relationship weights;

[0224] introduce blockchain technology to generate the blockchain hash value for the knowledge graph data G, and the hash formula is:

[0225] H G = Hash(G),

[0226] where H G represents the blockchain hash value of the knowledge graph, and Hash represents the hash function,

[0227] dynamically manage data access permissions using smart contracts, and the management formula is:

[0228] P c = Contract(G, A),

[0229] where P c represents the permission control policy, Contract is the smart contract, and A is the set of access permissions;

[0230] Specifically, a knowledge graph is constructed through semantic alignment to clearly describe cross-domain semantic relationships. Combining blockchain technology to record data flow and operations can improve data credibility and achieve dynamic management of access permissions.

[0231] Step S5: Embed the generated knowledge graph into the collaborative management platform and perform real-time analysis and dynamic optimization in combination with intelligent algorithms;

[0232] The intelligent algorithms include a dynamic optimization model, a prediction algorithm, and a resource scheduling algorithm;

[0233] The collaborative management platform displays the business status and data in real time through a visual dashboard and an analysis view;

[0234] The steps of performing real-time analysis and dynamic optimization in combination with intelligent algorithms are as follows:

[0235] Define the data set X and the prediction model P x :

[0236]

[0237] Among them, represents the prediction result, and P x represents the prediction model,

[0238] Define the objective function as J(x):

[0239] J(x) = g(x) - h(x),

[0240] Among them, J(x) is the objective function, g(x) represents the revenue function, and h(x) represents the cost function.

[0241] Solve the optimal solution through an optimization algorithm, and the formula is:

[0242] x * = arg max J(x),

[0243] Among them, x * represents the optimal solution of the objective function;

[0244] Specifically, the intelligent algorithm is used to predict the data trend in real time, improve the efficiency of the business process based on the dynamic optimization model, and combine with the visualization platform to intuitively present the analysis and optimization results to provide real-time support for decision-making.

[0245] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A multi-domain business collaborative management system based on blockchain, characterized by: include, The data collection module is responsible for collecting data from heterogeneous data sources in multiple fields, annotating data sources, timestamps, and operation records, and defining standardization rules, including field names, data types, and unique identifiers; The data cleaning module pre-processes the data collected by the data collection module, uses anomaly detection algorithms to remove invalid data, uses quality verification technology to perform consistency verification, integrity checks and duplicate data cleaning, and converts JSON, XML and CSV heterogeneous data formats into standardized formats through formatting tools; The data fusion module extracts feature data from preprocessed data, associates cross-domain data through modeling technology, and uses multimodal modeling technology to semantically align text, numerical, time series and image data to generate fused data; The trusted assurance module builds a knowledge graph based on fused data, performs semantic alignment, and introduces blockchain technology. The blockchain adopts a consortium chain structure and combines smart contracts to dynamically manage data access rights while recording the flow and operation of data. Collaborative management platform, used to receive knowledge graphs and manage businesses in multiple fields.

2. A multi-domain business collaborative management system based on blockchain as claimed in claim 1, characterized in that: In the collaborative management platform, intelligent algorithms are combined to optimize business processes in real time, and graph data is displayed through visual dashboards and analytical views.

3. A multi-domain business collaborative management method based on blockchain, based on a multi-domain business collaborative management system based on blockchain according to any one of claims 1 to 2, characterized in that: include: Step S1, collect data across fields, and mark the source, timestamp and operation record of the data during the collection process; During the data collection process, the field names, data types, and unique identifiers of the data are standardized; Step S2, preprocessing the data collected in step S1; Step S3, extracting features and correlating features of the data preprocessed in step S2 by using feature extraction and multimodal modeling technology to generate fused data; Step S4: Based on the fused data, the knowledge graph technology is used to generate semantic relationships between fields and construct a knowledge graph, and blockchain technology is introduced to record data flow and operations; Knowledge graphs enable modeling of cross-domain relationships through semantic alignment; The blockchain technology adopts a consortium chain structure and combines smart contracts to dynamically manage data access rights; Step S5, embedding the generated knowledge graph into the collaborative management platform, and combining it with intelligent algorithms for real-time analysis and dynamic optimization; Intelligent algorithms include dynamic optimization models, prediction algorithms, and resource scheduling algorithms; The collaborative management platform displays business status and data in real time through visual dashboards and analytical views.

4. A multi-domain business collaborative management method based on blockchain as claimed in claim 3, characterized in that: The step of standardizing the field name, data type and unique identifier of the data is as follows: Define the original collected data set as D: D = {d1, d2, ..., d n }, Where D represents the original collected data set, d i represents the i-th data entry, each data entry d i The composition form is: d i =(f i ,t i ,u i ), where f i Indicates the field name, t i Indicates the data type, u i Represents a unique identifier, Define the normalization rule set as R: Where R represents the set of standardized rules, Represents a standardized field name, Represents a standardized data type, Represents a standardized unique identifier; Apply the rule set R to the data set D to generate the standardized data set D s : Among them, D s represents the standardized data set, are the standardized field names, data types, and unique identifiers, respectively. Perform conflict detection, the detection function is: Among them, Q(D s ) represents the result of the conflict detection function, Indicates the conflicting field name pairs. Indicates the type of the conflicting field, A unique identifier representing the conflicting field. If Update the standardized rule R according to the conflict correction rule and regenerate D s , until 5. A multi-domain business collaborative management method based on blockchain as claimed in claim 4, characterized in that: The preprocessing includes: using anomaly detection algorithms and quality verification techniques to remove invalid data, and converting heterogeneous data into a standardized format through a formatting tool; The anomaly detection algorithm includes an analysis algorithm based on statistical distribution, an anomaly detection model based on machine learning, and a rule-driven boundary condition check; The quality verification technology includes consistency verification, integrity check and duplicate data cleaning; The formatting tool supports automatic mapping and standardization of JSON, XML, and CSV data formats.

6. A multi-domain business collaborative management method based on blockchain as claimed in claim 5, characterized in that: The steps of removing invalid data by using anomaly detection algorithm and quality verification technology are as follows: Assume that the data set is Z: Z = {z1,z2,…,z k }, Among them, Z represents the data set to be processed, z i represents the i-th data point in the data set, Calculate the probability density of the distribution. The probability density function is: Among them, P z (z) represents the distribution probability density of data point z, μ z represents the mean of the data set, σ z represents the standard deviation of the data set, exp represents the exponential function, Determine the set of outliers, and the determination formula is: WITH a ={z r ∣P z (With r )<∈ z }, Among them, Z a represents the set of outliers, z r represents the rth data point, ∈ z Indicates the threshold for outlier determination; Using the Isolation Forest Model F z Train the dataset Z and calculate the anomaly score for each data point using the following formula: Among them, S z (z t ) represents the data point z t The anomaly score, h z (z t ) represents the average split depth of the data point in the isolation forest, max(h z (Z)) represents the maximum value of all split depths in the data set, Determine abnormal data points, the determination formula is: WITH f ={z t ∣S z (With t )>τ z }, Among them, Z f represents the data set judged as abnormal, τ z represents the threshold of anomaly score, To perform consistency verification, define the field value set as V: V={v1,v2,…,v q }, Among them, V represents the value set of all data fields, v i Represents the i-th field value, Define the consistency rule function: V invalid ={v s ∣C v (v s )=False}, Among them, C v (v s ) represents the consistency rule function, V invalid Represents the set of all field values ​​that do not meet the consistency rules. False indicates the result that the field value does not meet the consistency rules. Define the data point similarity function: Among them, Sim z (z u ,z v ) represents the data point z u and z v The similarity of |z u -z v | represents the Euclidean distance between data points, max(|z u -z v |) indicates the maximum distance in the data set, If Sim z (z u ,z v )>λ z , then keep z u and delete z v .

7. A multi-domain business collaborative management method based on blockchain as claimed in claim 6, characterized in that: The feature extraction technology includes principal component analysis PCA based on dimensionality reduction, sparse coding and deep learning embedding model; The multimodal modeling adopts a data semantic alignment model to associate text, numerical, time series and image data to form fused data.

8. A multi-domain business collaborative management method based on blockchain as claimed in claim 7, characterized in that: The step of extracting and associating features from the data preprocessed in step S2 is as follows: Perform feature extraction and use principal component analysis PCA to reduce dimension, defining the data matrix as X: X=[x1,x2,…,x k ] T , Where X is an n×k data matrix containing k samples, x i represents the data vector of the i-th sample, T is the transposition operation, Calculate the covariance matrix, the calculation formula is: in, Σ X represents the covariance matrix of the data, μ X is the mean vector of the data, Perform eigendecomposition on the covariance matrix to obtain the principal components. The decomposition formula is: S X v j =λ j v j , Among them, v j is the jth eigenvector, λ j is the corresponding eigenvalue, Extract the first m principal components, and the data after dimensionality reduction is X′: X′=X·V m , Among them, X′ is the data matrix after dimension reduction, V m is the projection matrix consisting of the first m eigenvectors; Perform feature association and define multimodal data as M: M={T,N,S,I}, Among them, M represents a multimodal data set, T is text data, N is numerical data, S is time series data, and I is image data. Use semantic alignment function to associate multimodal data. The association formula is: R M =φ(T,N,S,I), Among them, R M is the semantic alignment result, φ represents the multimodal correlation function, The aligned data features are fused in the following way: F M =Concat(X T ,X N ,X S ,X I ), Among them, F M is the fused feature data, X T ,X N ,X S ,X I They are the feature representations of text, numerical value, time series and image respectively, and Concat represents the feature concatenation operation.

9. A multi-domain business collaborative management method based on blockchain as claimed in claim 8, characterized in that: The steps of using knowledge graph technology to generate semantic relationships between fields and construct knowledge graphs, and introducing blockchain technology to record data flow and operations are as follows: Construct a knowledge graph, define nodes and relationships, and define the node set of the knowledge graph as V: V={v1,v2,…,v m }, Among them, V is the node set, v i represents the i-th node, Define the relation set as E: E={e1,e2,…,e n }, Among them, E is the relationship set, e i represents the i-th relationship, The semantic alignment function is used to calculate the semantic weights between nodes. The calculation formula is: R ij =ψ(v i ,v j ), Among them, R ij Represents node v i and node v j The semantic relationship weight between them, ψ represents the semantic alignment function, Generate knowledge graph G: G=(V,E,R), Among them, G represents the knowledge graph, and R is the set of semantic relationship weights; Introduce blockchain technology to generate blockchain hash value for knowledge graph data G. The hash formula is: H G =Hash(G), Among them, H G Represents the blockchain hash value of the knowledge graph, Hash represents the hash function, Use smart contracts to dynamically manage data access permissions. The management formula is: P c =Contract(G,A), Among them, P c Represents the permission control strategy, Contract is the smart contract, and A is the access permission set.

10. A multi-domain business collaborative management method based on blockchain as claimed in claim 9, characterized in that: The steps of combining intelligent algorithms for real-time analysis and dynamic optimization are: Define the dataset X and the prediction model P x : in, Represents the prediction result, P x represents the prediction model, Define the objective function as J(x): J(x)=g(x)-h(x), Among them, J(x) is the objective function, g(x) represents the benefit function, and h(x) represents the cost function. The optimal solution is solved by the optimization algorithm. The formula is: x * =argmaxJ(x), Among them, x * represents the optimal solution of the objective function.

Images