A big data multi-scale fusion supervision system and method based on blockchain

Abstract

The present invention discloses a big data multi-scale fusion supervision system and method based on blockchain, which relates to the technical field of Internet data fusion supervision, extracts historical records of data on-chain information, collects records of on-chain information of various data, classifies the records, creates a data query index, analyzes use requests initiated by data users, extracts request contents in the use requests, classifies the extracted request contents, divides the request contents into main request contents and subordinate request contents, calculates the proportion of each main request content in the total request content, obtains word vectors of categories and word vectors of main request contents in the data query index, calculates the relative value of data to data use requests through the spatial distance relationship of word vectors, and calculates data information that meets transaction conditions as matching data for the request content.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet data fusion supervision, and specifically to a big data multi-scale fusion supervision system and method based on blockchain. Background Art

[0002] Blockchain technology uses block chain data structures to verify and store data, and uses distributed node consensus algorithms to generate and update data. Blockchain technology affects the deployment methods of the new generation of distributed cloud storage. The decentralization technology and encryption technology in the blockchain storage method require a large number of calculations through proprietary protocols and traversal on the blockchain, which requires users to have certain technical thresholds and technical conditions, which is not conducive to the flexible use of blockchain technology. Therefore, we need a blockchain data fusion management method that can perform search indexing to reduce unnecessary computing.

[0003] The Chinese patent with publication number CN112541019A and titled: Blockchain Resource Search Method and Device discloses a blockchain resource search method and device, which pre-calculates the attribute groups of all resources and the spatial attributes of all nodes, stores the attribute groups of resources on the node closest to the spatial location of the resource itself, assigns a start node to the searcher and sends the spatial attributes of the searcher to the start node. This resource search method requires a lot of computing power as a query retrieval support, and needs to be deployed on a cloud server, which is not conducive to flexible deployment. This method sets a high usage threshold and technical threshold for users. Users cannot directly search for content-related resources on the blockchain based on the content they need to search. Summary of the invention

[0004] The purpose of the present invention is to provide a big data multi-scale fusion supervision system and method based on blockchain to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a big data multi-scale fusion supervision system and method based on blockchain, the method comprising:

[0006] Step S100: In a blockchain for sharing data, extract the historical records of the on-chain information of the data, collect the records of the on-chain information of each piece of data, classify the records, and create a data query index;

[0007] Step S200: Analyze the use request initiated by the data user, extract the request content in the use request, classify the extracted request content, divide the request content into main request content and subordinate request content, and calculate the proportion of each main request content in the total request content;

[0008] Step S300: setting transaction weights of different main request contents according to the proportion of different contents in the total request contents and the relationship coefficient between the main request contents;

[0009] Step S400: Obtain the word vector of the category in the data query index and the word vector of the main request content, calculate the relative value of the data to the data use request through the spatial distance relationship of the word vectors, and calculate the data information that meets the transaction conditions;

[0010] Step S500: Calculate the correlation between the word vectors corresponding to the main request content and the correlation between the matching data of the main request content, and adjust the relationship coefficient between the main request content.

[0011] Furthermore, step S100 includes:

[0012] Step S101: extract the historical records of the data on-chain information, and record the block information and data tag information of each piece of data;

[0013] Step S102: obtaining a classification model of data labels, in which one data classification label corresponds to one node in the data classification model, and the association between data classification labels corresponds to the topological relationship of the nodes in the data label classification model;

[0014] Step S103: Create a data query index, where each entry in the index includes data category information, classification label information and block information.

[0015] Further, step S200 includes:

[0016] Step S201: Obtain a request content classification model of the data user's request content, in which a request content label corresponds to a node in the classification model, and the association between the request content labels corresponds to the topological relationship of the nodes in the request content classification model. The data user's request content is a data use request initiated by the user to the blockchain;

[0017] Step S202: extracting content features from the data user's request content, and classifying the extracted features into main request content and subordinate content corresponding to the main request content according to the content classification model, wherein the extracted content features correspond one-to-one to the nodes in the content classification model, and the content corresponding to the nodes that do not have the upper layer nodes among all the nodes is regarded as the main request content, and the content corresponding to all the lower layer classification nodes of the nodes corresponding to the main request content is the subordinate classification content corresponding to the main request content;

[0018] Step S203: extract the node information corresponding to the subordinate classification content corresponding to the main request content, extract the relative depth of the nodes corresponding to each subordinate classification content and the nodes corresponding to the main request content, the relative depth is the relative classification layer number of the nodes corresponding to the subordinate classification content and the nodes corresponding to the main request content, calculate the relative tendency value P of the main request content, wherein the relative tendency value is calculated as follows: P x It represents the relative tendency value of the xth main request content in a certain request content. The number of subordinate classification contents in the xth main request content is k, dep i Indicates the relative classification level of the node corresponding to the ith subordinate classification content and the node corresponding to the xth main request content among the k subordinate classification contents;

[0019] Step S204: Calculate the relative tendency value of each main request content, and calculate the proportion of each main request content in the total request content. The calculation method of the proportion of the main request content in the total request content is: Among them, the number of main request contents in a certain request content is t, α e Indicates the proportion of the e-th main request content in a request content to the total request content, P e represents the relative tendency value of the e-th main request content, P j represents the relative tendency value of the jth main request content among the t main request contents;

[0020] A request contains several main request contents. The data user will further describe and limit each main request content. The more detailed and accurate the main request content is, the higher its relative tendency value in the requested content will be. The relative tendency value is the basis for arranging priorities during the data matching process.

[0021] Furthermore, step S300 includes:

[0022] Step S301: sorting the main request contents from high to low according to the proportion of each main request content in the total request content;

[0023] Step S302: Calculate the transaction weight allocation coefficient λ corresponding to the main request content, where the transaction weight allocation coefficient λ of the e-th main request content in a certain request content is e ,λ e The calculation method is λ e =f e α e , f e The algebraic sum of the relationship coefficients between the main request content whose ranking is lower than the e-th main request content in a certain request content and the e-th main request content, α eIndicates the proportion of the e-th main request content in the total request content;

[0024] Step S303: Set the transaction weight available for a data use request initiated by the data user to G, and calculate the transaction weight of each main request content. In a certain request content, the transaction weight corresponding to the e-th main request content is g e , g e The calculation method is: g e =λ e ·G.

[0025] Furthermore, the step of calculating the data information satisfying the transaction conditions in step S400 includes:

[0026] Step S401: clustering the word vectors represented by the data category information and classification label information of the data owner through a clustering algorithm, wherein the data category information is used as the core of the clustering, and the clustering radius of each cluster is r;

[0027] Step S402: Calculate the distance between the word vector representing the main request content and each cluster core, select the cluster core with the smallest distance as the target cluster core, the distance between the word vector of the main request content and the corresponding target cluster core is the difference distance, calculate the size between the difference distance and the cluster radius of the corresponding cluster,

[0028] Step S403: When d2<r, calculate The calculation result is the first value evaluation coefficient, where d2 is the difference distance between the word vector representing the main request content and the corresponding target cluster core, d1 represents the distance between the word vector corresponding to the classification label of the data and the corresponding cluster core, and d3 is the distance between the word vector representing the main request content and the word vector corresponding to the classification label;

[0029] When d2 ≥ r, calculate The calculation result of is the second value evaluation coefficient;

[0030] Step S404: Calculate the transaction weight of the data according to the relationship between the calculated difference distance and the cluster radius of the corresponding cluster, h = μ1·G or h = μ2·G, where h represents the relative value of the data to the data use request, μ1 represents the first value evaluation coefficient, and μ2 represents the second value evaluation coefficient;

[0031] Step S405: Adjust the position of the selected data to match the data using the data request, when h≤g e Under the condition of e The matching data corresponding to the main request content, g eThe transaction weight corresponding to the e-th main request content in a request content is calculated. According to the arrangement order of the main request content, data matching is performed on each main request content in turn, and the matched data is traded with the blockchain corresponding to the data.

[0032] Furthermore, step S500 includes:

[0033] Step S501: respectively calculating the correlation between the word vectors corresponding to the main request content and the correlation between the matching data of the main request content;

[0034] Step S502: setting a first correction threshold γ1 and a second correction threshold γ2, wherein the first correction threshold is used to correct overfitting of matching data, and the second correction threshold is used to correct missing correlation of matching data;

[0035] Step S503: When S(a1, b1)>S(a1 * , b1 * ), and S(a1, b1)-S(a1 * , b1 * )>γ1, the relationship coefficient between the main request contents is adjusted as follows: f1'=q1·(S(a1, b1)-S(a1 * , b1 * ))f1, where a1 represents the word vector corresponding to a main request content c1, b1 represents the word vector corresponding to a main request content c2 that ranks lower than c1 in the main request content sorting sequence, a1 * Indicates the feature vector corresponding to the data matched by c1, b1 * represents the feature vector corresponding to the data matched by c2, S(a1, b1) represents the correlation between a1 and b1, S(a1 * , b1 * ) indicates a1 * With b1 * The correlation degree, f1 represents the initial value of the relationship coefficient between c1 and c2, f1' represents the value after the initial value of the relationship coefficient between c1 and c2 is adjusted, f1' is used as the next relationship coefficient between c1 and c2, and q1 is the first adjustment coefficient;

[0036] Step S504: When S(a1, b1)<S(a1 * , b1 * ), and S(a1 * , b1 * )-S(a1, b1)<γ2, the relationship coefficient adjustment method between the main request contents is, f1'=q2·((S(a1 * , b1 *)-S(a1, b1)) / S(a1, b1))·f1, f1' represents the value after the initial value of the relationship coefficient between c1 and c2 is adjusted, f1' is used as the relationship coefficient between c1 and c2 for the next time, and q2 is the second adjustment coefficient.

[0037] In order to better implement the above method, a big data multi-scale fusion supervision system based on blockchain is also proposed. The system includes: a data query index management module, a data usage request analysis module, a data transaction supervision module and a relationship coefficient adjustment module. The data query index management module is used to create and store data query indexes of blockchain, wherein the data usage request analysis module is used to analyze the content of data user requests, the data transaction supervision module is used to manage data transactions, and the relationship coefficient adjustment module is used to adjust the relationship coefficients between the main request contents.

[0038] Furthermore, the data usage request analysis module includes: a request content feature extraction unit, a node depth calculation unit, a relative tendency value calculation unit and a relative tendency value sorting unit, wherein the content feature extraction unit is used to extract content features of the data user's request content, the node depth calculation unit is used to calculate the relative depth of the node corresponding to the request content in the classification model, the relative tendency value calculation unit is used to calculate the relative tendency value of the main request content in the request content, and the relative tendency value sorting unit is used to sort the main request content according to the relative tendency value.

[0039] Furthermore, the data transaction supervision module includes a transaction weight allocation unit, a cluster core selection unit, a difference distance comparison unit and a transaction data matching unit, wherein the transaction weight allocation unit is used to allocate transaction weights to the main request content, the cluster core selection unit is used to select the cluster core for the main request content, the difference distance comparison unit is used to calculate the difference distance between the main request content and the cluster core, and the transaction data matching unit is used to match data for the main request content.

[0040] Furthermore, the relationship coefficient adjustment module includes: a request content correlation calculation unit, a matching data correlation calculation unit, a correction judgment unit and a relationship coefficient adjustment unit, wherein the main request content correlation calculation unit is used to calculate the correlation between the main request contents, the correlation calculation unit is used to calculate the correlation between the main request contents and the matching data, the correction judgment unit is used to judge whether the relationship coefficient adjustment needs to be adjusted, and the relationship coefficient adjustment unit is used to adjust the relationship coefficient between the main request contents.

[0041] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: the present invention creates a data query index, collects the block information and data tag information provided by the data owner, and obtains the data query index. Through the index, on the basis of ensuring the original confidentiality of the data, the advantage of making the data query directional is added. By analyzing the request content initiated by the data user, the tendency of the request content is analyzed to provide a basis for data search. By comparing the relationship between the request content and the relationship between the request content matching data, the relationship coefficient between the main request content is adaptively adjusted, so that the matched data and the request content maintain a certain degree of correlation, which not only reduces the lack of correlation, but also controls overfitting, and further improves the matching data to meet the usage needs of the data user. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0043] Figure 1 This is a schematic diagram of the structure of a big data multi-scale fusion supervision system based on blockchain in the present invention patent;

[0044] Figure 2 This is a flow chart of a big data multi-scale fusion supervision method based on blockchain in the patent of this invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] See also Figure 1 and Figure 2 , the present invention provides a technical solution:

[0047] Step S100: In a blockchain for sharing data, extract the historical records of the on-chain information of the data, collect the records of the on-chain information of each piece of data, classify the records, and create a data query index;

[0048] Wherein, step S100 includes:

[0049] Step S101: extract the historical records of the data on-chain information, and record the block information and data tag information of each piece of data;

[0050] Step S102: obtaining a classification model of data labels, in which one data classification label corresponds to one node in the data classification model, and the association between data classification labels corresponds to the topological relationship of the nodes in the data label classification model;

[0051] Step S103: Create a data query index, where each entry in the index includes data category information, classification label information and block information.

[0052] Step S200: Analyze the use request initiated by the data user, extract the request content in the use request, classify the extracted request content, divide the request content into main request content and subordinate request content, and calculate the proportion of each main request content in the total request content;

[0053] Wherein, step S200 includes:

[0054] Step S201: Obtain a request content classification model of the data user's request content, in which a request content label corresponds to a node in the classification model, and the association between the request content labels corresponds to the topological relationship of the nodes in the request content classification model. The data user's request content is a data use request initiated by the user to the blockchain;

[0055] Step S202: extracting content features from the data user's request content, and classifying the extracted features into main request content and subordinate content corresponding to the main request content according to the content classification model, wherein the extracted content features correspond one-to-one to the nodes in the content classification model, and the content corresponding to the nodes that do not have the upper layer nodes among all the nodes is regarded as the main request content, and the content corresponding to all the lower layer classification nodes of the nodes corresponding to the main request content is the subordinate classification content corresponding to the main request content;

[0056] Step S203: extract the node information corresponding to the subordinate classification content corresponding to the main request content, extract the relative depth of the nodes corresponding to each subordinate classification content and the nodes corresponding to the main request content, the relative depth is the relative classification layer number of the nodes corresponding to the subordinate classification content and the nodes corresponding to the main request content, calculate the relative tendency value P of the main request content, wherein the relative tendency value is calculated as follows: P x It represents the relative tendency value of the xth main request content in a certain request content. The number of subordinate classification contents in the xth main request content is k, dep i Indicates the relative classification level of the node corresponding to the ith subordinate classification content and the node corresponding to the xth main request content among the k subordinate classification contents;

[0057] Step S204: Calculate the relative tendency value of each main request content, and calculate the proportion of each main request content in the total request content. The calculation method of the proportion of the main request content in the total request content is: Among them, the number of main request contents in a certain request content is t, α e Indicates the proportion of the e-th main request content in a request content to the total request content, P e represents the relative tendency value of the e-th main request content, P j Indicates the relative tendency value of the jth main request content among t main request contents.

[0058] Step S300: setting transaction weights of different main request contents according to the proportion of different contents in the total request contents and the relationship coefficient between the main request contents;

[0059] Wherein, step S300 includes:

[0060] Step S301: sorting the main request contents from high to low according to the proportion of each main request content in the total request content;

[0061] Step S302: Calculate the transaction weight allocation coefficient λ corresponding to the main request content, where the transaction weight allocation coefficient λ of the e-th main request content in a certain request content is e ,λ e The calculation method is λ e =f e α e , f e The algebraic sum of the relationship coefficients between the main request content whose ranking is lower than the e-th main request content in a certain request content and the e-th main request content, α e Indicates the proportion of the e-th main request content in the total request content;

[0062] Step S303: Set the transaction weight available for a data use request initiated by the data user to G, and calculate the transaction weight of each main request content. In a certain request content, the transaction weight corresponding to the e-th main request content is g e , g e The calculation method is: g e =λ e ·G.

[0063] Step S400: Obtain the word vector of the category in the data query index and the word vector of the main request content, calculate the relative value of the data to the data use request through the spatial distance relationship of the word vectors, and calculate the data information that meets the transaction conditions;

[0064] Methods for training and obtaining word vectors: For example, training word vectors through the Word2vec model, using Python language, calling the gensim library, jieba library, and re library to implement Word2vec model training, and outputting similar words and their cosine distances;

[0065] Commonly used models for obtaining word vectors include: GloVe model and FastText model;

[0066] For example, word vector visualization methods include: using the t-SNE algorithm to reduce the dimension, and calling matplotlib to plot the vector results to achieve word vector visualization;

[0067] The step of calculating the data information satisfying the transaction conditions in step S400 includes:

[0068] Step S401: clustering the word vectors represented by the data category information and classification label information of the data owner through a clustering algorithm, wherein the data category information is used as the core of the clustering, and the clustering radius of each cluster is r;

[0069] Step S402: Calculate the distance between the word vector representing the main request content and each cluster core, select the cluster core with the smallest distance as the target cluster core, the distance between the word vector of the main request content and the corresponding target cluster core is the difference distance, calculate the size between the difference distance and the cluster radius of the corresponding cluster,

[0070] Step S403: When d2<r, calculate The calculation result is the first value evaluation coefficient, where d2 is the difference distance between the word vector representing the main request content and the corresponding target cluster core, d1 represents the distance between the word vector corresponding to the classification label of the data and the corresponding cluster core, and d3 is the distance between the word vector representing the main request content and the word vector corresponding to the classification label;

[0071] When d2 ≥ r, calculate The calculation result of is the second value evaluation coefficient;

[0072] Step S404: Calculate the transaction weight of the data according to the relationship between the calculated difference distance and the cluster radius of the corresponding cluster, h = μ1·G or h = μ2·G, where h represents the relative value of the data to the data use request, μ1 represents the first value evaluation coefficient, and μ2 represents the second value evaluation coefficient;

[0073] Step S405: Adjust the position of the selected data to match the data using the data request, when h≤g e Under the condition of e The matching data corresponding to the main request content, g eThe transaction weight corresponding to the e-th main request content in a request content is calculated. According to the arrangement order of the main request content, data matching is performed on each main request content in turn, and the matched data is traded with the blockchain corresponding to the data.

[0074] Step S500: Calculate the correlation between the word vectors corresponding to the main request content and the correlation between the matching data of the main request content, and adjust the relationship coefficient between the main request content;

[0075] Wherein, step S500 includes:

[0076] Step S501: respectively calculating the correlation between the word vectors corresponding to the main request content and the correlation between the matching data of the main request content;

[0077] Step S502: setting a first correction threshold γ1 and a second correction threshold γ2, wherein the first correction threshold is used to correct overfitting of matching data, and the second correction threshold is used to correct missing correlation of matching data;

[0078] Step S503: When S(a1, b1)>S(a1 * , b1 * ), and S(a1, b1)-S(a1 * , b1 * )>γ1, the relationship coefficient between the main request contents is adjusted as follows: f1'=q1·(S(a1, b1)-S(a1 * , b1 * ))f1, where a1 represents the word vector corresponding to a main request content c1, b1 represents the word vector corresponding to a main request content c2 that ranks lower than c1 in the main request content sorting sequence, a1 * Indicates the feature vector corresponding to the data matched by c1, b1 * represents the feature vector corresponding to the data matched by c2, S(a1, b1) represents the correlation between a1 and b1, S(a1 * , b1 * ) indicates a1 * With b1 * The correlation degree, f1 represents the initial value of the relationship coefficient between c1 and c2, f1' represents the value after the initial value of the relationship coefficient between c1 and c2 is adjusted, f1' is used as the next relationship coefficient between c1 and c2, and q1 is the first adjustment coefficient;

[0079] Step S504: When S(a1, b1)<S(a1 * , b1 * ), and S(a1 * , b1 *)-S(a1, b1)<γ2, the relationship coefficient adjustment method between the main request contents is, f1'=q2·((S(a1 * , b1 * )-S(a1, b1)) / S(a1, b1))·f1, f1' represents the value after the initial value of the relationship coefficient between c1 and c2 is adjusted, f1' is used as the relationship coefficient between c1 and c2 for the next time, and q2 is the second adjustment coefficient.

[0080] Methods for calculating associations include: calculating Euclidean distance, Manhattan distance, cosine similarity or Pearson correlation coefficient. In order to measure the uniformity of association, when calculating S(a1, b1) and S(a1 * , b1 * ) using the same correlation calculation method.

[0081] The system includes: a data query index management module, a data usage request analysis module, a data transaction supervision module and a relationship coefficient adjustment module. The data query index management module is used to create and store data query indexes of the blockchain, wherein the data usage request analysis module is used to analyze the content of data user requests, the data transaction supervision module is used to manage data transactions, and the relationship coefficient adjustment module is used to adjust the relationship coefficients between the main request contents.

[0082] Among them, the data usage request analysis module includes: a request content feature extraction unit, a node depth calculation unit, a relative tendency value calculation unit and a relative tendency value sorting unit, wherein the content feature extraction unit is used to extract content features of the data user's request content, the node depth calculation unit is used to calculate the relative depth of the node corresponding to the request content in the classification model, the relative tendency value calculation unit is used to calculate the relative tendency value of the main request content in the request content, and the relative tendency value sorting unit is used to sort the main request content according to the relative tendency value.

[0083] Among them, the data transaction supervision module includes a transaction weight allocation unit, a cluster core selection unit, a difference distance comparison unit and a transaction data matching unit, wherein the transaction weight allocation unit is used to allocate transaction weights to the main request content, the cluster core selection unit is used to select the cluster core for the main request content, the difference distance comparison unit is used to calculate the difference distance between the main request content and the cluster core, and the transaction data matching unit is used to match data for the main request content.

[0084] Among them, the relationship coefficient adjustment module includes: a request content correlation calculation unit, a matching data correlation calculation unit, a correction judgment unit and a relationship coefficient adjustment unit, among which the main request content correlation calculation unit is used to calculate the correlation between the main request contents, the correlation calculation unit is used to calculate the correlation between the main request contents and the matching data, the correction judgment unit is used to judge whether the relationship coefficient adjustment needs to be adjusted, and the relationship coefficient adjustment unit is used to adjust the relationship coefficient between the main request contents.

[0085] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0086] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A big data multi-scale fusion supervision method based on blockchain, characterized in that: The method comprises the following steps: Step S100: In a blockchain for sharing data, extract the historical records of the on-chain information of the data, collect the records of the on-chain information of each piece of data, classify the records, and create a data query index; Step S200: Analyze the use request initiated by the data user, extract the request content in the use request, classify the extracted request content, divide the request content into main request content and subordinate request content, and calculate the proportion of each main request content in the total request content; Step S200 includes: Step S201: Obtain a request content classification model of the data user's request content, in which a request content label corresponds to a node in the classification model, and the association between the request content labels corresponds to the topological relationship of the nodes in the request content classification model. The data user's request content is a data use request initiated by the user to the blockchain; Step S202: extracting content features from the data user's request content, and classifying the extracted features into main request content and subordinate content corresponding to the main request content according to the content classification model, wherein the extracted content features correspond one-to-one to the nodes in the content classification model, and the content corresponding to the nodes that do not have the upper layer nodes among all the nodes is regarded as the main request content, and the content corresponding to all the lower layer classification nodes of the nodes corresponding to the main request content is the subordinate classification content corresponding to the main request content; Step S203: extract the node information corresponding to the subordinate classification content corresponding to the main request content, extract the relative depth of the nodes corresponding to each subordinate classification content and the nodes corresponding to the main request content, the relative depth is the relative classification layer number of the nodes corresponding to the subordinate classification content and the nodes corresponding to the main request content, calculate the relative tendency value P of the main request content, wherein the relative tendency value is calculated as follows: P x It represents the relative tendency value of the xth main request content in a certain request content. The number of subordinate classification contents in the xth main request content is k, dep i Indicates the relative classification level of the node corresponding to the ith subordinate classification content and the node corresponding to the xth main request content among the k subordinate classification contents; Step S204: Calculate the relative tendency value of each main request content, and calculate the proportion of each main request content in the total request content. The calculation method of the proportion of the main request content in the total request content is: Among them, the number of main request contents in a certain request content is t, α e Indicates the proportion of the e-th main request content in a request content to the total request content, P e represents the relative tendency value of the e-th main request content, P j Represents the relative tendency value of the jth main request content among t main request contents Step S300: setting transaction weights of different main request contents according to the proportion of different contents in the total request contents and the relationship coefficient between the main request contents; Step S400: Obtain the word vector of the category in the data query index and the word vector of the main request content, calculate the relative value of the data to the data use request through the spatial distance relationship of the word vectors, and calculate the data information that meets the transaction conditions; Step S500: Calculate the correlation between the word vectors corresponding to the main request content and the correlation between the matching data of the main request content, and adjust the relationship coefficient between the main request content; Step S500 includes: Step S501: respectively calculating the correlation between the word vectors corresponding to the main request content and the correlation between the matching data of the main request content; Step S502: setting a first correction threshold γ1 and a second correction threshold γ2, wherein the first correction threshold is used to correct overfitting of matching data, and the second correction threshold is used to correct missing correlation of matching data; Step S503: When S(a1, b1)>S(a1 * , b1 * ), and S(a1, b1)-S(a1 * , b1 * )>γ1, the relationship coefficient between the main request contents is adjusted as follows: f1'=q1·(S(a1, b1)-S(a1 * , b1 * ))f1, where a1 represents the word vector corresponding to a main request content c1, b1 represents the word vector corresponding to a main request content c2 that ranks lower than c1 in the main request content sorting sequence, a1 * Indicates the feature vector corresponding to the data matched by c1, b1 * represents the feature vector corresponding to the data matched by c2, S(a1, b1) represents the correlation between a1 and b1, S(a1 * , b1 * ) means a1 * With b1 * The correlation degree, f1 represents the initial value of the relationship coefficient between c1 and c2, f1' represents the value after the initial value of the relationship coefficient between c1 and c2 is adjusted, f1' is used as the next relationship coefficient between c1 and c2, and q1 is the first adjustment coefficient; Step S504: When S(a1, b1)<S(a1 * , b1 * ), and S(a1 * , b1 * )-S(a1, b1)<γ2, the relationship coefficient adjustment method between the main request contents is, f1'=q2·((S(a1 * , b1 * )-S(a1, b1)) / S(a1, b1))·f1, f1' represents the value after the initial value of the relationship coefficient between c1 and c2 is adjusted, f1' is used as the relationship coefficient between c1 and c2 for the next time, and q2 is the second adjustment coefficient.

2. According to a blockchain-based big data multi-scale fusion supervision method according to claim 1, it is characterized in that: Step S100 includes: Step S101: extract the historical records of the data on-chain information, and record the block information and data tag information of each piece of data; Step S102: obtaining a classification model of data labels, in which one data classification label corresponds to one node in the data classification model, and the association between data classification labels corresponds to the topological relationship of the nodes in the data label classification model; Step S103: Create a data query index, where each entry in the index includes data category information, classification label information and block information.

3. According to a blockchain-based big data multi-scale fusion supervision method according to claim 2, it is characterized in that: The step S300 includes: Step S301: sorting the main request contents from high to low according to the proportion of each main request content in the total request content; Step S302: Calculate the transaction weight allocation coefficient λ corresponding to the main request content, where the transaction weight allocation coefficient λ of the e-th main request content in a certain request content is e ,λ e The calculation method is λ e =f e α e , f e The algebraic sum of the relationship coefficients between the main request content whose ranking is lower than the e-th main request content in a certain request content and the e-th main request content, α e Indicates the proportion of the e-th main request content in the total request content; Step S303: Set the transaction weight available for a data use request initiated by the data user to G, and calculate the transaction weight of each main request content. In a certain request content, the transaction weight corresponding to the e-th main request content is g e , g e The calculation method is: g e =λ e ·G.

4. According to a blockchain-based big data multi-scale fusion supervision method according to claim 3, it is characterized in that: The step of calculating the data information satisfying the transaction conditions in step S400 includes: Step S401: clustering the word vectors represented by the data category information and classification label information of the data owner through a clustering algorithm, wherein the data category information is used as the core of the clustering, and the clustering radius of each cluster is r; Step S402: Calculate the distance between the word vector representing the main request content and each cluster core, select the cluster core with the smallest distance as the target cluster core, the distance between the word vector of the main request content and the corresponding target cluster core is the difference distance, calculate the size between the difference distance and the cluster radius of the corresponding cluster, Step S403: When d2<r, calculate The calculation result is the first value evaluation coefficient, where d2 is the difference distance between the word vector representing the main request content and the corresponding target cluster core, d1 represents the distance between the word vector corresponding to the classification label of the data and the corresponding cluster core, and d3 is the distance between the word vector representing the main request content and the word vector corresponding to the classification label; When d2 ≥ r, calculate The calculation result of is the second value evaluation coefficient; Step S404: Calculate the transaction weight of the data according to the relationship between the calculated difference distance and the cluster radius of the corresponding cluster, h = μ1·G or h = μ2·G, where h represents the relative value of the data to the data use request, μ1 represents the first value evaluation coefficient, and μ2 represents the second value evaluation coefficient; Step S405: Adjust the position of the selected data to match the data using the data request, when h≤g e Under the condition of e The matching data corresponding to the main request content, g e The transaction weight corresponding to the e-th main request content in a request content is calculated. According to the arrangement order of the main request content, data matching is performed on each main request content in turn, and the matched data is traded with the blockchain corresponding to the data.

5. A big data multi-scale fusion supervision system applied to a big data multi-scale fusion supervision method based on blockchain as described in any one of claims 1 to 4, characterized in that: The system includes: a data query index management module, a data usage request analysis module, a data transaction supervision module and a relationship coefficient adjustment module. The data query index management module is used to create and store data query indexes of blockchains, wherein the data usage request analysis module is used to analyze the content of data user requests, the data transaction supervision module is used to manage data transactions, and the relationship coefficient adjustment module is used to adjust the relationship coefficients between the main request contents.

6. The big data multi-scale fusion supervision system according to claim 5 is characterized in that: The data usage request analysis module includes: a request content feature extraction unit, a node depth calculation unit, a relative tendency value calculation unit and a relative tendency value sorting unit, wherein the content feature extraction unit is used to extract content features of the data user's request content, the node depth calculation unit is used to calculate the relative depth of the node corresponding to the request content in the classification model, the relative tendency value calculation unit is used to calculate the relative tendency value of the main request content in the request content, and the relative tendency value sorting unit is used to sort the main request content according to the relative tendency value.

7. The big data multi-scale fusion supervision system according to claim 6 is characterized in that: The data transaction supervision module includes a transaction weight allocation unit, a cluster core selection unit, a difference distance comparison unit and a transaction data matching unit, wherein the transaction weight allocation unit is used to allocate transaction weights to the main request content, the cluster core selection unit is used to select the cluster core for the main request content, the difference distance comparison unit is used to calculate the difference distance between the main request content and the cluster core, and the transaction data matching unit is used to match data for the main request content.

8. The big data multi-scale fusion supervision system according to claim 7 is characterized in that: The relationship coefficient adjustment module includes: a request content correlation calculation unit, a matching data correlation calculation unit, a correction judgment unit and a relationship coefficient adjustment unit, wherein the main request content correlation calculation unit is used to calculate the correlation between the main request contents, the correlation calculation unit is used to calculate the correlation between the main request contents and the matching data, the correction judgment unit is used to judge whether the relationship coefficient adjustment needs to be adjusted, and the relationship coefficient adjustment unit is used to adjust the relationship coefficient between the main request contents.

Images