Supply chain financing method and system based on data analysis

Abstract

The invention belongs to the technical field of supply chain financing, and discloses a supply chain financing method and system based on data analysis. The method comprises the following steps: building a cloud data center based on a supply chain financing platform, and constructing a heterogeneous data cleaning model and a supply chain financing evaluation model; performing heterogeneous data cleaning by using a heterogeneous data cleaning model according to the plurality of pieces of real-time bid-winning data; acquiring real-time financial data according to the plurality of standard real-time bid-winning data, and performing supply chain financing evaluation by using the supply chain financing evaluation model according to the plurality of real-time financial data; generating a real-time supply chain financing application report, and sending the report to a financial institution; and receiving the real-time authorization limit data of the supply chain enterprise returned by the financial institution, and generating a real-time supply chain financing project. According to the invention, the problems of long approval process, low application efficiency, low evaluation efficiency, poor accuracy and large risk in the prior art are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of supply chain financing, and specifically relates to a supply chain financing method and system based on data analysis. Background Art

[0002] Supply chain financing is a management method that optimizes the cash flow of the supply chain through financing, which can reduce the financing cost in the supply chain and improve the overall operational efficiency of the supply chain. In particular, after the supply chain enterprises have won the bid for certain large projects, they have the problem of tight funds. They need to submit the winning bid information, the basic financial situation of the enterprise and other data to the supply chain financing platform. After the supply chain financing platform collects the above data, it will conduct financing evaluation, apply for credit from financial institutions, and finally finance according to the credit line applied for.

[0003] Existing supply chain financing methods have long approval processes and low application efficiency, which delays the financing progress of supply chain companies and affects the normal progress of projects. They usually rely on manual financing assessments, which have problems such as low assessment efficiency and poor accuracy. They are unable to discover the potential financing risks of supply chain companies, resulting in greater financial risks for supply chain financing platforms. Summary of the invention

[0004] In order to solve the problems of long approval process, low application efficiency, low evaluation efficiency, poor accuracy and great risks in the prior art, the present invention aims to provide a supply chain financing method and system based on data analysis.

[0005] The technical solution adopted by the present invention is:

[0006] A supply chain financing method based on data analysis includes the following steps:

[0007] Based on the supply chain financing platform, a cloud data center is built, and a heterogeneous data cleaning model and a supply chain financing evaluation model are constructed in the cloud data center;

[0008] Collect some real-time winning bid data from the bidding and procurement platform, upload some real-time winning bid data to the cloud data center, and use the heterogeneous data cleaning model to clean the heterogeneous data according to some real-time winning bid data to obtain some standard real-time winning bid data in a unified format;

[0009] According to certain standard real-time winning bid data, collect the corresponding real-time financial data of supply chain enterprises, upload certain real-time financial data to the cloud data center, and use the supply chain financing evaluation model to conduct supply chain financing evaluation based on certain real-time financial data to obtain certain real-time supply chain financing evaluation results;

[0010] Generate a corresponding real-time supply chain financing application report based on the standard real-time bidding data, real-time financial data and real-time supply chain financing evaluation results of supply chain enterprises, and send it to financial institutions;

[0011] Receive real-time authorization quota data of supply chain enterprises returned by financial institutions and generate real-time supply chain financing projects.

[0012] Furthermore, based on the supply chain financing platform, a cloud data center is built, and a heterogeneous data cleaning model and a supply chain financing evaluation model are constructed in the cloud data center, including the following steps:

[0013] Based on the supply chain financing platform, a cloud data center is built to collect historical bidding data and historical financial data of several supply chain companies and send them to the cloud data center;

[0014] Based on the cloud data center, data preprocessing is performed on a number of historical winning bid data and a number of historical financial data to obtain a number of preprocessed historical winning bid data and a number of preprocessed historical financial data;

[0015] Based on several pre-processed historical bid winning data, a reinforcement learning algorithm is used to build a heterogeneous data cleaning model;

[0016] Based on several pre-processed historical financial data, a supply chain financing evaluation model is constructed using a deep learning algorithm.

[0017] Furthermore, based on several pre-processed historical bid winning data, a reinforcement learning algorithm is used to construct a heterogeneous data cleaning model, which includes the following steps:

[0018] Build a unified data model based on the differences in data structures of pre-processed historical bid-winning data from different supply chain companies;

[0019] According to the unified data model, set the corresponding historical data mapping and conversion strategy for each pre-processed historical winning bid data;

[0020] Based on several pre-processed historical bidding data and the corresponding historical data mapping and conversion strategies, a heterogeneous data cleaning model is constructed using the reinforcement learning algorithm.

[0021] Furthermore, the heterogeneous data cleaning model is constructed based on the DQN algorithm.

[0022] Furthermore, based on some pre-processed historical financial data, a supply chain financing evaluation model is constructed using a deep learning algorithm, including the following steps:

[0023] Performing data dimensionality reduction on a number of pre-processed historical financial data to obtain a number of reduced-dimensional historical financial data and a set of key financial indicators;

[0024] Labeling is performed on a number of historical financial data after dimensionality reduction to obtain a number of historical financial data after dimensionality reduction with real supply chain financing evaluation labels;

[0025] Based on several dimensionality-reduced historical financial data with real supply chain financing evaluation labels, a supply chain financing evaluation model is constructed using a deep learning algorithm.

[0026] Furthermore, the supply chain financing evaluation model is constructed based on the IPCO-N-GAN-BiLSTM-Attention algorithm, where N is the total number of attention dimensions of financial data.

[0027] Furthermore, a number of real-time winning bid data from the bidding and procurement platform are collected, uploaded to the cloud data center, and a heterogeneous data cleaning model is used to clean the heterogeneous data according to the real-time winning bid data to obtain a number of standard real-time winning bid data in a unified format, including the following steps:

[0028] Collecting a number of real-time winning bid data from the bidding and procurement platform, uploading the real-time winning bid data to the cloud data center, and preprocessing the real-time winning bid data based on the cloud data center to obtain a number of preprocessed real-time winning bid data;

[0029] Based on the pre-processed real-time winning bid data, use the heterogeneous data cleaning model to generate the corresponding real-time data mapping and conversion strategy;

[0030] According to the real-time data mapping and conversion strategy, the pre-processed real-time winning bid data is mapped and converted to obtain standard real-time winning bid data;

[0031] All pre-processed real-time winning bid data are traversed to obtain a number of standard real-time winning bid data in a unified data model format.

[0032] Furthermore, according to a number of standard real-time winning bid data, real-time financial data of corresponding supply chain enterprises are collected, the real-time financial data are uploaded to a cloud data center, and supply chain financing evaluation is performed based on the real-time financial data using a supply chain financing evaluation model to obtain a number of real-time supply chain financing evaluation results, including the following steps:

[0033] According to the standard real-time winning bid data, connect to the corresponding supply chain enterprise server, collect the real-time financial data of the supply chain enterprise stored in the supply chain enterprise server, and upload some real-time financial data to the cloud data center;

[0034] Based on the cloud data center, data preprocessing is performed on a number of real-time financial data to obtain a number of preprocessed real-time financial data, and data dimension reduction is performed on the preprocessed real-time financial data according to a set of key financial indicators to obtain a number of reduced-dimensional real-time financial data;

[0035] Use the supply chain financing evaluation model to extract N real-time attention dimension features of the real-time financial data after dimensionality reduction, and obtain the real-time attention weight of each real-time attention dimension feature;

[0036] According to the N real-time attention dimension characteristics and the real-time attention weights, the real-time financial comprehensive data characteristics are obtained, and the supply chain financing evaluation is performed according to the real-time financial comprehensive data characteristics to obtain the real-time supply chain financing evaluation results;

[0037] Traverse all the real-time financial data after dimensionality reduction and obtain several real-time supply chain financing evaluation results.

[0038] Furthermore, the real-time financial data includes real-time trade data, real-time financing repayment behavior data, real-time financing application records and real-time financial data of supply chain enterprises over a period of time.

[0039] A supply chain financing system based on data analysis is used to implement a supply chain financing method. The system includes a supply chain financing platform and a cloud data center. The supply chain financing platform is communicatively connected to the cloud data center, and the supply chain financing platform is also communicatively connected to an external procurement platform, financial institutions, and several supply chain enterprise servers. The cloud data center is provided with a heterogeneous data cleaning model and a supply chain financing evaluation model.

[0040] The beneficial effects of the present invention are:

[0041] The present invention discloses a supply chain financing method and system based on data analysis. By building a cloud data center and conducting cloud supply chain financing evaluation, the automation, intelligence and evaluation efficiency of the supply chain financing evaluation are improved; a heterogeneous data cleaning model is constructed to clean the heterogeneous winning bid data of different bidding and procurement platforms to ensure the uniform format and improve the efficiency and accuracy of subsequent data analysis; the winning bid information of supply chain enterprises is automatically collected to provide data support for subsequent financing projects, solving the problems of long approval process and low efficiency, and financing can be obtained in the winning bid stage, which solves the capital shortage problem of enterprises to a certain extent and improves the stability of the enterprise's supply chain; by constructing a supply chain financing evaluation model, accurate and efficient data analysis is performed on the financial data of supply chain enterprises, which can timely discover potential financing risks and adjust corresponding financing strategies, thereby reducing the risk level of the supply chain financing platform.

[0042] Other beneficial effects of the present invention will be further described in the specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a flowchart of the supply chain financing method based on data analysis in the present invention.

[0044] Figure 2It is a structural block diagram of the supply chain financing system based on data analysis in the present invention. DETAILED DESCRIPTION

[0045] The present invention will be further explained below in conjunction with the accompanying drawings and specific embodiments.

[0046] Embodiment 1:

[0047] like Figure 1 As shown, this embodiment provides a supply chain financing method based on data analysis, comprising the following steps:

[0048] S1: Based on the supply chain financing platform, a cloud data center is built, and a heterogeneous data cleaning model and a supply chain financing evaluation model are constructed in the cloud data center, including the following steps:

[0049] S1-1: Based on the supply chain financing platform, a cloud data center is built to collect historical bidding data and historical financial data of several supply chain companies and send them to the cloud data center;

[0050] S1-2: Based on the cloud data center, data preprocessing is performed on a number of historical winning bid data and a number of historical financial data to obtain a number of preprocessed historical winning bid data and a number of preprocessed historical financial data;

[0051] Data preprocessing includes data cleaning, duplicate elimination, and normalization in order to improve data quality and scale the data to a uniform scale to improve the speed and effect of model training;

[0052] S1-3: Based on some pre-processed historical bid winning data, use the DQN algorithm to build a heterogeneous data cleaning model, including the following steps:

[0053] S1-3-1: Based on the differences in the data structures of the pre-processed historical bid winning data of different supply chain enterprises, a unified data model is constructed, including the following steps:

[0054] S1-3-1-1: Analyze the pre-processed historical bid winning data of different bidding and procurement platforms to obtain several core data elements of each pre-processed historical bid winning data;

[0055] Since the data systems used by various bidding and procurement platforms are inconsistent, the historical bid winning data come from different data sources and have different data formats. This embodiment analyzes the data format characteristics in each bidding and procurement platform system and identifies the core data elements across systems.

[0056] S1-3-1-2: According to several core data elements of different procurement platforms, set the data structure of the unified data model, such as tables, fields, relations, etc.;

[0057] S1-3-1-3: According to the data structure of the unified data model, set the data relationship between core data elements, including one-to-one, one-to-many, many-to-many and other relationships;

[0058] S1-3-1-4: According to the core data elements of different procurement platforms, the data structure of the unified data model and the data relationship between the core data elements, set the data constraints of each core data element, such as data type, value range, default value, etc.;

[0059] S1-3-1-5: Construct a corresponding unified data model based on the core data elements of different procurement platforms, the data structure of the unified data model, the data relationship between the core data elements, and the data constraints of each core data element;

[0060] The unified data model serves as the basis for cleaning heterogeneous data. It covers the core data elements in the data systems of various procurement platforms and serves as a model support for cleaning heterogeneous data into winning bid data in a unified format.

[0061] S1-3-2: According to the unified data model, set the corresponding historical data mapping and conversion strategy for each pre-processed historical winning bid data;

[0062] Data mapping and conversion strategies include mapping strategies and data conversion strategies. Mapping strategies are used to determine how to map heterogeneous data into a unified data model. Data conversion strategies are used to convert heterogeneous data into data in a unified data model based on a preset conversion algorithm or script. Data mapping and conversion strategies are used to handle data heterogeneity and resolve inconsistencies in data formats, data types, data semantics, etc. in different data sources.

[0063] S1-3-3: Based on several pre-processed historical bid winning data and corresponding historical data mapping and conversion strategies, a heterogeneous data cleaning model is constructed using the Deep Q-network (DQN) algorithm, including the following steps:

[0064] S1-3-3-1: According to the unified data model and several historical data mapping and conversion strategies, set the data mapping and conversion goals of the heterogeneous data cleaning model, and define the simulation scenarios of the heterogeneous data cleaning model according to the data mapping and conversion goals;

[0065] S1-3-3-2: Use the DQN algorithm to build a deep Q network and Q function, agency, and experience storage pool for heterogeneous data cleaning models;

[0066] S1-3-3-3: According to the data mapping and conversion goals and simulation scenarios, define the state space S = [s1,...,s i ,...,sI ] and action space A=[a1,...,a j ,...,a I' ], where s i is the i-th state value, i is the state indicator, I is the size of the state space, a j is the j-th action value, j is the action indicator, I' is the size of the action space, and according to the state space S = [s1,...,s i ,...,s I ] and action space A=[a1,...,a j ,...,a I' ], setting the reward function R(s) of the heterogeneous data cleaning model p ,a p ,s' p ), where s p is the pth state value, p is the state indicator, a p is the pth action value, p is the action indication, s' p is the pth state value at the next moment;

[0067] S1-3-3-4: Based on the state space S = [s1, ..., s i ,...,s I ], action space A = [a1,...,a j ,...,a I' ] and reward function R(s p ,a p ,s' p ), using some pre-processed historical bid winning data and corresponding historical data mapping and conversion strategies, optimizing the deep Q network and agency to obtain the optimal deep Q network and the optimal agency, and generating some historical data mapping and conversion experiences;

[0068] S1-3-3-5: Based on the optimal deep Q network and the optimal agency, a heterogeneous data cleaning model is constructed, and several historical data mapping and conversion experiences are stored in the experience storage pool;

[0069] S1-4: Based on some pre-processed historical financial data, the Improved Crested Porcupine Optimizer (ICPO)-N-Generative Adversarial Network (GAN)-Bidirectional Long Short-Term Memory (BiLSTM)-Attention algorithm is used to build a supply chain financing evaluation model, including the following steps:

[0070] S1-4-1: Use the principal component analysis (PCA) method to reduce the dimension of some pre-processed historical financial data to obtain some reduced-dimensional historical financial data and key financial indicator sets;

[0071] Data dimensionality reduction reduces the data volume, improves the efficiency of subsequent model training, and enhances the ability of data features to represent important information;

[0072] S1-4-2: Labeling a number of historical financial data after dimensionality reduction to obtain a number of historical financial data after dimensionality reduction with real supply chain financing evaluation labels;

[0073] S1-4-3: Divide a number of historical financial data with real supply chain financing evaluation labels after dimensionality reduction into a model training set and a model test set in a ratio of 7:3;

[0074] S1-4-4: Use the IPCO-N-GAN-BiLSTM-Attention algorithm to build an initial supply chain financing assessment model, where N is the total number of financial data focus dimensions, which include corporate financial situation, corporate financing credit situation, corporate operating situation, and corporate industry status. Through the above dimensions, the financing risks and tolerance of supply chain companies can be better assessed;

[0075] The supply chain financing evaluation model includes a model network parameter optimization module based on the IPCO algorithm, N attention dimension feature extraction modules based on the GAN algorithm, a supply chain financing evaluation prediction module based on the BiLSTM algorithm, and an attention setting module based on the Attention mechanism, including the following steps:

[0076] S1-4-4-1: Use the first initial network parameters of the GAN network and the second initial network parameters of the BiLSTM network as the optimization targets of the IPCO algorithm;

[0077] S1-4-4-2: Based on the optimization goal of the IPCO algorithm, use the model network parameter optimization module to optimize the model network parameters to obtain the optimal first initial network parameters and the optimal second initial network parameters, including the following steps:

[0078] S1-4-4-2-1: Encode the first initial network parameters of the GAN network and the second initial network parameters of the BiLSTM network as the positions of the IPCO individuals of the IPCO optimization algorithm;

[0079] S1-4-4-2-2: Set the IPCO population parameters, maximum number of iterations and fitness function of the IPCO optimization algorithm;

[0080] S1-4-4-2-3: According to the number of individuals in the IPCO population parameters, the Circle chaotic mapping sequence is used for initialization to obtain the initial IPCO population. The formula is:

[0081]

[0082] In the formula, is the initial IPCO individual of the Circle chaos map; is the randomly generated initial IPCO individual; l' is the IPCO individual indicator;

[0083] S1-4-4-2-4: Introduce a cyclic population reduction mechanism to limit the number of individuals in the IPCO population parameters and obtain the updated IPCO population parameters for the next iteration. The formula is:

[0084]

[0085] In the formula, S t+1 is the number of individuals in the IPCO population parameter of the t+1th iteration; S t is the number of individuals in the IPCO population parameter of the tth iteration; S min is the minimum number of individuals in the IPCO population parameter; a is the function evaluation parameter; V is the function evaluation cycle parameter; V max Evaluate loop parameters for the maximum function;

[0086] S1-4-4-2-5: According to the fitness function, calculate the initial fitness value of the initial IPCO individual in the initial IPCO population. The formula is:

[0087]

[0088] In the formula, For IPCO individuals The fitness value; MSE is the prediction mean square error function; Y' l' , Y l' IPCO individuals The corresponding predicted and true values; S is the number of individuals in the IPCO population parameter;

[0089] S1-4-4-2-6: according to the initial fitness value and the updated IPCO population parameters, the first defense strategy, the second defense strategy, the third defense strategy and the fourth defense strategy are used to update the initial IPCO population to obtain an updated IPCO population;

[0090] The formula for the first defense strategy is:

[0091]

[0092] In the formula, Updated IPCO individuals within the first defense scope; is the initial IPCO individual within the first defense range; τ1 is a random number based on normal distribution; τ2 is a random value in the interval [0,1]; It is the optimal solution within the first defense range; is the vector generated between the true optimal solution within the first defense range and the optimal solution randomly selected from the IPCO population; l' is the IPCO individual indicator; t is the iteration indicator;

[0093] The formula for the second defense strategy is:

[0094]

[0095] In the formula, Updated IPCO individuals within the second defense range; It is the initial IPCO individual within the second defense range; is the search upper limit vector of the second defense range; τ3 is a random value in the interval [0,1]; are the r1th and r2th initial IPCO individuals respectively; r1 and r2 are two random integers between [1, S]; is the vector generated between the true optimal solution within the second defense range and the optimal solution randomly selected from the IPCO population;

[0096] The formula for the third defense strategy is:

[0097]

[0098] In the formula, Updated IPCO individuals within the third defense scope; It is the initial IPCO individual within the third defense range; is the search upper limit vector of the third defense range; are the r2th and r3th initial IPCO individuals respectively; r3 is a random integer between [1, S]; The odor diffusion factor defined for the fitness function; λ t It is a defense factor; Control parameters for search direction;

[0099] The formula for the fourth defense strategy is:

[0100]

[0101] In the formula, Updated IPCO individuals within the fourth defense scope; It is the initial IPCO individual within the fourth defense range; is the optimal solution within the fourth defense range; τ4 and τ5 are both random values ​​in the interval [0,1]; λ t It is a defense factor; Control parameters for search direction; is the average force affecting the search direction; a' is the convergence speed factor;

[0102] S1-4-4-2-7: Use the dynamic reverse learning algorithm to perform dynamic reverse learning on the updated IPCO population to generate a dynamic reverse IPCO population. The formula is:

[0103]

[0104] In the formula, is the dynamic reverse IPCO individual; γ is the decreasing inertia coefficient; L max , L min are the maximum and minimum values ​​of the vector space respectively; For the updated IPCO individual;

[0105] S1-4-4-2-8: According to the fitness function, calculate the fitness values ​​of all IPCO individuals in the updated IPCO population and the dynamically reversed IPCO population, and take the IPCO individual with the minimum fitness value as the optimal individual;

[0106] S1-4-4-2-9: If the number of iterations reaches the maximum number of iterations or the fitness value of the optimal individual meets the requirements, the optimal solution corresponding to the current optimal individual is output to obtain the optimal first initial network parameters of the GAN network and the optimal second initial network parameters of the BiLSTM network;

[0107] The IPCO optimization algorithm can accurately optimize the initial network structure of the GAN network and the BiLSTM network, which is very important for the initialization of network parameters, because a good initialization can improve the convergence speed and final performance of the algorithm. The IPCO optimization algorithm can find the global optimal solution to the integer programming problem, which means that the best initial parameters can be found for the GAN network and the BiLSTM network in all possible parameter configurations. By optimizing the initial network parameters of the GAN network and the BiLSTM network, the generalization ability of the model can be improved, and the risk of overfitting or underfitting can be reduced, thereby obtaining better performance on unseen data.

[0108] S1-4-4-3: Based on the optimal first initial network parameters, use the GAN algorithm to construct N initial focus dimension feature extraction modules;

[0109] S1-4-4-4: Based on the optimal second initial network parameters, use the BiLSTM algorithm to build the initial supply chain financing evaluation prediction module;

[0110] S1-4-4-5: Based on the Attention mechanism, an attention setting module is constructed between the N initial attention dimension feature extraction modules and the initial supply chain financing evaluation prediction module to obtain the initial supply chain financing evaluation model;

[0111] S1-4-5: Based on the model training set, the initial supply chain financing evaluation model is optimized and trained to obtain an optimized supply chain financing evaluation model;

[0112] S1-4-6: Based on the model test set, the optimized supply chain financing evaluation model is tested to obtain several historical prediction supply chain financing evaluation labels;

[0113] S1-4-7: Statistically compare several historically predicted supply chain financing evaluation labels with corresponding real supply chain financing evaluation labels to obtain the corresponding model test accuracy;

[0114] S1-4-8: If the model test accuracy is greater than the accuracy threshold, the optimal supply chain financing evaluation model is output, otherwise, the optimization training continues;

[0115] S2: Collect some real-time winning bid data from the bidding and procurement platform, upload some real-time winning bid data to the cloud data center, and use the heterogeneous data cleaning model to clean the heterogeneous data according to the real-time winning bid data to obtain some standard real-time winning bid data in a unified format, including the following steps:

[0116] S2-1: Collect some real-time winning bid data from the bidding and procurement platform, upload some real-time winning bid data to the cloud data center, and pre-process some real-time winning bid data based on the cloud data center to obtain some pre-processed real-time winning bid data;

[0117] S2-2: Based on the pre-processed real-time winning bid data, use the heterogeneous data cleaning model to generate the corresponding real-time data mapping and conversion strategy, including the following steps:

[0118] S2-2-1: Based on the pre-processed real-time winning bid data, update the action space and state space of the strategy generation model to obtain the updated action space A'=[a'1,...,a' j" ,...,a' I' ] and the updated state space S'=[s'1,...,s'i" ,...,s' I ], where s' i" is the i-th state value after update, i is the state indicator, a' j" is the jth "action value" after update, j" is the action indicator;

[0119] S2-2-2: Input real-time heterogeneous data into the strategy generation model and update the action space A'=[a'1,...,a' j" ,...,a' I' ] and the updated state space S'=[s'1,...,s' i" ,...,s' I ], using the Q function of the deep Q network, we get the predicted Q value of the possible action;

[0120] S2-2-3: Use the agent to take the possible action corresponding to the highest predicted Q value as the execution action, and obtain the reward value R'(s) for the execution action according to the reward function p' ,a p' ,s' p' );

[0121] S2-2-4: Update the predicted Q value of the possible action according to the reward value of the executed action, obtain the updated Q value of the possible action, and repeat the previous step until the iteration number threshold is reached;

[0122] The formula is:

[0123] Q(s' p' ,a' p' )=(1-α)·Q(s p' ,a p' )+α·(R(s p' ,a p' ,s' p' )+γ·Q max (s p' ,a p' ))

[0124] In the formula, Q(s' p' ,a' p' ) is the updated state value s' p' and the updated action value a' p' The corresponding updated Q value; Q(s p' ,a p' ) is the state value s p' and action value a p' The corresponding predicted Q value; α is the learning rate; Q max (s p' ,a p' ) is the highest predicted Q value;

[0125] S2-2-5: Based on the updated Q value, a greedy strategy is used to output the execution action corresponding to the highest updated Q value as a real-time data mapping and conversion strategy;

[0126] S2-3: According to the real-time data mapping and conversion strategy, the pre-processed real-time winning bid data is mapped and converted to obtain standard real-time winning bid data;

[0127] S2-4: traverse all pre-processed real-time winning bid data to obtain a number of standard real-time winning bid data in the same format as the unified data model;

[0128] S3: Based on certain standard real-time winning bid data, collect the corresponding real-time financial data of the supply chain enterprises, upload certain real-time financial data to the cloud data center, and use the supply chain financing evaluation model to conduct supply chain financing evaluation based on certain real-time financial data to obtain certain real-time supply chain financing evaluation results, including the following steps:

[0129] S3-1: Connect to the corresponding supply chain enterprise server according to the standard real-time winning bid data, collect the real-time financial data of the supply chain enterprise stored in the supply chain enterprise server, and upload some real-time financial data to the cloud data center;

[0130] Real-time financial data includes real-time trade data, real-time financing repayment behavior data, real-time financing application records, and real-time financial data of supply chain companies over a period of time;

[0131] S3-2: Based on the cloud data center, data preprocessing is performed on a number of real-time financial data to obtain a number of preprocessed real-time financial data, and data dimension reduction is performed on the number of preprocessed real-time financial data according to a set of key financial indicators to obtain a number of reduced-dimensional real-time financial data;

[0132] S3-3: Use the supply chain financing evaluation model to extract N real-time attention dimension features of the real-time financial data after dimensionality reduction, and obtain the real-time attention weight of each real-time attention dimension feature;

[0133] S3-4: Based on the N real-time attention dimension features and the real-time attention weights, the real-time financial comprehensive data features are obtained, and supply chain financing evaluation is performed based on the real-time financial comprehensive data features to obtain the real-time supply chain financing evaluation results;

[0134] S3-5: Traverse all the real-time financial data after dimensionality reduction to obtain several real-time supply chain financing evaluation results;

[0135] S4: Generate a corresponding real-time supply chain financing application report based on the standard real-time bidding data, real-time financial data and real-time supply chain financing evaluation results of the supply chain enterprises, and send it to the financial institution;

[0136] S5: Receive the real-time authorization quota data of the supply chain enterprises returned by the financial institutions and generate real-time supply chain financing projects.

[0137] Embodiment 2:

[0138] like Figure 2 As shown, this embodiment provides a supply chain financing system based on data analysis, which is used to implement a supply chain financing method. The system includes a supply chain financing platform and a cloud data center. The supply chain financing platform is connected to the cloud data center in communication, and the supply chain financing platform is also connected to an external procurement platform, a financial institution, and a plurality of supply chain enterprise servers in communication. The cloud data center is provided with a heterogeneous data cleaning model and a supply chain financing evaluation model.

[0139] The supply chain financing platform is used to build a cloud data center, and to construct a heterogeneous data cleaning model and a supply chain financing evaluation model in the cloud data center; to collect a number of real-time winning bid data from the bidding and procurement platform, and upload the real-time winning bid data to the cloud data center; to collect the real-time financial data of the supply chain enterprises stored in the corresponding supply chain enterprise servers based on a number of standard real-time winning bid data, and upload the real-time financial data to the cloud data center; to generate a corresponding real-time supply chain financing application report based on the standard real-time winning bid data, real-time financial data and real-time supply chain financing evaluation results of the supply chain enterprises, and send it to the financial institution; to receive the real-time authorization quota data of the supply chain enterprises returned by the financial institution, and generate a real-time supply chain financing project;

[0140] The cloud data center is used to clean heterogeneous data based on a number of real-time winning bid data using a heterogeneous data cleaning model to obtain a number of standard real-time winning bid data in a unified format; and to conduct supply chain financing evaluation based on a number of real-time financial data using a supply chain financing evaluation model to obtain a number of real-time supply chain financing evaluation results.

[0141] The present invention discloses a supply chain financing method and system based on data analysis. By building a cloud data center and conducting cloud supply chain financing evaluation, the automation, intelligence and evaluation efficiency of the supply chain financing evaluation are improved; a heterogeneous data cleaning model is constructed to clean the heterogeneous winning bid data of different bidding and procurement platforms to ensure the uniform format and improve the efficiency and accuracy of subsequent data analysis; the winning bid information of supply chain enterprises is automatically collected to provide data support for subsequent financing projects, solving the problems of long approval process and low efficiency, and financing can be obtained in the winning bid stage, which solves the capital shortage problem of enterprises to a certain extent and improves the stability of the enterprise's supply chain; by constructing a supply chain financing evaluation model, accurate and efficient data analysis is performed on the financial data of supply chain enterprises, which can timely discover potential financing risks and adjust corresponding financing strategies, thereby reducing the risk level of the supply chain financing platform.

[0142] The present invention is not limited to the above optional implementations, and anyone can derive other various forms of products under the enlightenment of the present invention. The above specific implementations should not be understood as limiting the scope of protection of the present invention. The scope of protection of the present invention should be based on the definition in the claims, and the description can be used to interpret the claims.

Claims

1. A supply chain financing method based on data analysis, characterized by: The steps include: Based on the supply chain financing platform, a cloud data center is built, and a heterogeneous data cleaning model and a supply chain financing evaluation model are constructed in the cloud data center; Collect some real-time winning bid data from the bidding and procurement platform, upload some real-time winning bid data to the cloud data center, and use the heterogeneous data cleaning model to clean the heterogeneous data according to some real-time winning bid data to obtain some standard real-time winning bid data in a unified format; According to certain standard real-time winning bid data, collect the corresponding real-time financial data of supply chain enterprises, upload certain real-time financial data to the cloud data center, and use the supply chain financing evaluation model to conduct supply chain financing evaluation based on certain real-time financial data to obtain certain real-time supply chain financing evaluation results; Generate a corresponding real-time supply chain financing application report based on the standard real-time bidding data, real-time financial data and real-time supply chain financing evaluation results of supply chain enterprises, and send it to financial institutions; Receive real-time authorization quota data of supply chain enterprises returned by financial institutions and generate real-time supply chain financing projects.

2. A supply chain financing method based on data analysis according to claim 1, characterized in that: Based on the supply chain financing platform, a cloud data center is built, and a heterogeneous data cleaning model and a supply chain financing evaluation model are constructed in the cloud data center, including the following steps: Based on the supply chain financing platform, a cloud data center is built to collect historical bidding data and historical financial data of several supply chain companies and send them to the cloud data center; Based on the cloud data center, data preprocessing is performed on a number of historical winning bid data and a number of historical financial data to obtain a number of preprocessed historical winning bid data and a number of preprocessed historical financial data; Based on several pre-processed historical bid winning data, a reinforcement learning algorithm is used to build a heterogeneous data cleaning model; Based on several pre-processed historical financial data, a supply chain financing evaluation model is constructed using a deep learning algorithm.

3. A supply chain financing method based on data analysis according to claim 2, characterized in that: Based on several pre-processed historical bid winning data, a reinforcement learning algorithm is used to build a heterogeneous data cleaning model, which includes the following steps: Build a unified data model based on the differences in data structures of pre-processed historical bid-winning data from different supply chain companies; According to the unified data model, set the corresponding historical data mapping and conversion strategy for each pre-processed historical winning bid data; Based on several pre-processed historical bidding data and the corresponding historical data mapping and conversion strategies, a heterogeneous data cleaning model is constructed using the reinforcement learning algorithm.

4. A supply chain financing method based on data analysis according to claim 3, characterized in that: The heterogeneous data cleaning model is built based on the DQN algorithm.

5. A supply chain financing method based on data analysis according to claim 2, characterized in that: Based on some pre-processed historical financial data, a supply chain financing evaluation model is constructed using a deep learning algorithm, including the following steps: Performing data dimensionality reduction on a number of pre-processed historical financial data to obtain a number of reduced-dimensional historical financial data and a set of key financial indicators; Labeling is performed on a number of historical financial data after dimensionality reduction to obtain a number of historical financial data after dimensionality reduction with real supply chain financing evaluation labels; Based on several dimensionality-reduced historical financial data with real supply chain financing evaluation labels, a supply chain financing evaluation model is constructed using a deep learning algorithm.

6. A supply chain financing method based on data analysis according to claim 5, characterized in that: The supply chain financing evaluation model is constructed based on the IPCO-N-GAN-BiLSTM-Attention algorithm, where N is the total number of attention dimensions of financial data.

7. A supply chain financing method based on data analysis according to claim 4, characterized in that: Collect some real-time winning bid data from the bidding and procurement platform, upload some real-time winning bid data to the cloud data center, and use the heterogeneous data cleaning model to clean the heterogeneous data according to some real-time winning bid data to obtain some standard real-time winning bid data in a unified format, including the following steps: Collecting a number of real-time winning bid data from the bidding and procurement platform, uploading the real-time winning bid data to the cloud data center, and preprocessing the real-time winning bid data based on the cloud data center to obtain a number of preprocessed real-time winning bid data; Based on the pre-processed real-time winning bid data, use the heterogeneous data cleaning model to generate the corresponding real-time data mapping and conversion strategy; According to the real-time data mapping and conversion strategy, the pre-processed real-time winning bid data is mapped and converted to obtain standard real-time winning bid data; All pre-processed real-time winning bid data are traversed to obtain a number of standard real-time winning bid data in a unified data model format.

8. A supply chain financing method based on data analysis according to claim 6, characterized in that: According to certain standard real-time winning bid data, real-time financial data of corresponding supply chain enterprises are collected, and certain real-time financial data are uploaded to the cloud data center. Based on certain real-time financial data, supply chain financing evaluation models are used to conduct supply chain financing evaluation, and certain real-time supply chain financing evaluation results are obtained, including the following steps: According to the standard real-time winning bid data, connect to the corresponding supply chain enterprise server, collect the real-time financial data of the supply chain enterprise stored in the supply chain enterprise server, and upload some real-time financial data to the cloud data center; Based on the cloud data center, data preprocessing is performed on a number of real-time financial data to obtain a number of preprocessed real-time financial data, and data dimension reduction is performed on the preprocessed real-time financial data according to a set of key financial indicators to obtain a number of reduced-dimensional real-time financial data; Use the supply chain financing evaluation model to extract N real-time attention dimension features of the real-time financial data after dimensionality reduction, and obtain the real-time attention weight of each real-time attention dimension feature; According to the N real-time attention dimension characteristics and the real-time attention weights, the real-time financial comprehensive data characteristics are obtained, and the supply chain financing evaluation is performed according to the real-time financial comprehensive data characteristics to obtain the real-time supply chain financing evaluation results; Traverse all the real-time financial data after dimensionality reduction and obtain several real-time supply chain financing evaluation results.

9. The supply chain financing method based on data analysis according to claim 1, characterized in that: The real-time financial data includes real-time trade data, real-time financing repayment behavior data, real-time financing application records and real-time financial data of supply chain enterprises over a period of time.

10. A supply chain financing system based on data analysis, used to implement the supply chain financing method according to any one of claims 1 to 9, characterized in that: The system includes a supply chain financing platform and a cloud data center. The supply chain financing platform is communicatively connected to the cloud data center, and the supply chain financing platform is also communicatively connected to an external procurement platform, financial institutions, and several supply chain enterprise servers. The cloud data center is provided with a heterogeneous data cleaning model and a supply chain financing evaluation model.

Images