Data processing method based on multi-channel financing

Abstract

The invention relates to the technical field of data processing, in particular to a data processing method based on multi-channel financing, and the method comprises the steps: obtaining a financial data sequence, a credit sequence and a financing scheme sequence of an enterprise sample; clustering the enterprise samples according to the financial data sequence and the credit sequence to obtain a plurality of enterprise clusters; after orderly sample clustering is carried out on the financing scheme average sequence in the enterprise cluster, influence coefficients of financial data and credits in the enterprise cluster on the financing scheme are calculated respectively; and in the enterprise cluster to which the target enterprise belongs, calculating a second similarity between the target enterprise and each enterprise sample in the historical financing period according to the influence coefficient, and carrying out weighted summation on financing schemes of each enterprise sample in the historical financing period to obtain a recommended financing scheme. According to the technical scheme, the financing scheme of the target enterprise can be accurately obtained.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a data processing method based on multi-channel financing. Background Art

[0002] Enterprise financing refers to the process of financing funds with the enterprise as the main body, so that the supply and demand of funds between the enterprise and its internal links change from imbalance to balance. In the financing process, enterprises can obtain financing services through different financing channels such as banks, securities companies or investment funds. There are differences in the risks and financing costs of different financing channels. How to select the financing channel suitable for the enterprise and provide the enterprise with personalized financing solutions is an urgent problem to be solved.

[0003] At present, a patent application document with publication number CN118656730A discloses a financing plan generation method and system based on supply chain data, wherein the method includes: obtaining first information and second information, the first information includes basic information of enterprises applying for financial services, and the second information includes basic information of enterprises in the supply chain, supply chain transaction data, logistics information and financial institution data; performing evaluation processing based on the first information and the second information to obtain a supply chain influence scoring result; performing data fusion processing on the second information to obtain a supply chain fusion data set; performing identification processing based on the supply chain fusion data set to obtain a supply chain network model; performing data integration processing on the basic information of the enterprise, the supply chain influence scoring result and the supply chain network model based on a graph neural network mathematical model to obtain a comprehensive data view; performing time series analysis processing based on the comprehensive data view to obtain an evaluation result; and optimizing the evaluation result based on a deep learning mathematical model to obtain a personalized financing plan.

[0004] The above method integrates multi-source heterogeneous data in the supply chain, obtains the complex relationships between enterprises in the supply chain, and then obtains personalized financing plans for enterprises. However, the enterprises in the supply chain are upstream and downstream relationships in the industrial chain. There are large differences in the business areas, financial conditions and information records between the enterprises. These differences will lead to inaccurate generated personalized financing plans. Summary of the invention

[0005] In order to solve the technical problem of inaccurate financing plans, this application provides a data processing method based on multi-channel financing, which can accurately obtain the financing plan of the target enterprise.

[0006] In a first aspect, the present application provides a data processing method based on multi-channel financing, the processing method comprising: obtaining a financial data sequence, a credit sequence and a financing plan sequence of multiple enterprise samples within a preset time period, the preset time period including multiple historical financing cycles, and the financing plan is the proportion of financing volume of each financing channel within a financing cycle; clustering the enterprise samples according to the financial data sequence and the credit sequence to obtain multiple enterprise clusters; performing ordered sample clustering on the average sequence of financing plans within the enterprise cluster, dividing the preset time period into multiple time sub-segments, and respectively calculating the influence coefficients of the financial data and credit in the enterprise cluster on the financing plan; the influence coefficient of the financial data is negatively correlated with the variance of the financial data in the time sub-segment, and is positively correlated with the DTW distance of the financial data between adjacent time sub-segments; in the enterprise cluster to which the target enterprise belongs, calculating the second similarity between the target enterprise in the current financing cycle and each enterprise sample in the historical financing cycle according to the influence coefficient of the financial data and the credit, and weighted summing the financing plans of each enterprise sample in the historical financing cycle according to the second similarity to obtain a recommended financing plan for the target enterprise.

[0007] Enterprises with the same industry as the target enterprise and for which financing plan sequences can be collected are taken as enterprise samples. The enterprise samples are divided into multiple enterprise clusters according to the financial data sequence and credit sequence within the preset time period, and one enterprise cluster corresponds to one enterprise category; further, the average sequence of financing plans within an enterprise cluster is clustered in an ordered sample, and the preset time period is divided into multiple time sub-segments; the financing plans within each time sub-segment are basically the same, and the differences in financing plans between adjacent time sub-segments are large. When the variance of financial data in each time sub-segment is large, it means that changes in financial data will not cause changes in financing plans, and financial data have little impact on financing plans; between two adjacent time sub-segments, if the financial data between the two adjacent time sub-segments is The DTW distance is small, indicating that the financial data between two adjacent time sub-segments has not changed, that is, the change in the financing plan is not caused by the change in the financial data, and the impact of the financial data is small, and then the influence coefficient of the financial data and credit on the financing plan in the enterprise cluster can be calculated; in the enterprise cluster to which the target enterprise belongs, the second similarity between the target enterprise in the current financing cycle and the enterprise samples in the historical financing cycle is calculated based on the influence coefficient, and the financing plans of the enterprise samples in the historical financing cycle are weighted and summed according to the second similarity to obtain the recommended financing plan of the target enterprise, which can eliminate the financing plan errors caused by differences in business areas, financial conditions and information records between enterprises, and accurately obtain the financing plan of the target enterprise.

[0008] Preferably, the financial data includes at least one of liabilities, working capital, expenditures and income; and the credit rating is the default rate within a financing cycle.

[0009] Preferably, obtaining multiple enterprise clusters includes: clustering enterprise samples using a Kmeans clustering algorithm, determining the number of clusters using an elbow method, and obtaining multiple enterprise clusters, wherein the clustering distance between any enterprise samples in the Kmeans clustering algorithm is the sum of the DTW distance of the financial data sequence and the DTW distance of the credit sequence.

[0010] The enterprise samples are classified unsupervisedly and divided into multiple enterprise clusters. One enterprise cluster corresponds to one category, and the financial data and credit rating of each enterprise sample in one enterprise cluster are basically the same.

[0011] Preferably, the ordered sample clustering of the average sequence of financing plans within the enterprise cluster includes: performing ordered sample clustering on the average sequence of financing plans according to the initial number to obtain the clustering results of the initial number, the clustering results including the initial number of initial time sub-segments; calculating the evaluation value of the initial number, the evaluation value is negatively correlated with the variance of the average financing plan in the initial time sub-segment, and is positively correlated with the DTW distance of the adjacent initial time sub-segments; updating the initial number to draw the evaluation value curve, obtaining the clustering results of the initial number corresponding to the inflection point in the evaluation value curve, and obtaining multiple time sub-segments.

[0012] The preset time period is divided into multiple time sub-segments according to the average sequence of financing plans. The financing plans of each financing cycle in a time sub-segment are basically the same, while the financing plans are quite different between adjacent time sub-segments. At the same time, the clustering results of the initial number of inflection points in the evaluation value curve are used as multiple time sub-segments, which can prevent the problem of too short time sub-segments due to too large initial number, and provide a data basis for calculating the impact coefficient.

[0013] Preferably, before respectively calculating the influence coefficients of the financial data and credit ratings in the enterprise cluster on the financing scheme, the processing method further comprises: deleting time sub-segments whose length is less than a preset length.

[0014] When the number of historical financing cycles in a time sub-segment is less than the preset length, it means that the time sub-segment is noise data, and the time sub-segment is deleted to avoid the noise data affecting the calculation of subsequent impact coefficients and improve the accuracy of the impact coefficients.

[0015] Preferably, the enterprise cluster The impact coefficient of financial data for: , For time sub-segment and time sub-segments The DTW distance of financial data, is the number of time sub-segments, For time sub-segment The variance of financial data.

[0016] Preferably, the method for obtaining the enterprise cluster to which the target enterprise belongs includes: collecting the financial data sequence and credit sequence of the target enterprise within a preset time period before the current financing cycle; calculating the first similarity between the target enterprise and each enterprise cluster, and taking the enterprise cluster corresponding to the maximum sequence similarity as the enterprise cluster to which the target enterprise belongs; the first similarity is the sequence similarity between the financial data sequence and the average sequence of financial data in the enterprise cluster, and the average of the sequence similarity between the credit sequence and the average sequence of credit in the enterprise cluster.

[0017] Preferably, the target enterprise has a historical financing cycle The second similarity of enterprise sample c for: ; and is the influence coefficient of financial data and creditworthiness in the enterprise cluster to which the target enterprise belongs, and The target companies and historical financing cycles are The distance similarity of financial data and creditworthiness in the internal enterprise sample c.

[0018] Preferably, obtaining the recommended financing plan for the target enterprise includes: performing weighted summation of the financing plans of each enterprise sample in the historical financing cycle using the normalized second similarity to obtain the recommended financing plan for the target enterprise.

[0019] Preferably, before obtaining the recommended financing plan for the target enterprise, the processing method further comprises: deleting enterprise samples within multiple historical financing cycles whose second similarity is not greater than a similarity threshold.

[0020] In order to avoid the influence of financing plans whose second similarity is less than the similarity threshold on the recommended financing plan, the recommended financing plan of the target enterprise is calculated based on the financing plans corresponding to the enterprise samples in the historical financing cycle whose second similarity is greater than the similarity threshold, so as to accurately obtain the financing plan of the target enterprise.

[0021] The technical solution of this application has the following beneficial technical effects: Enterprises with the same industry as the target enterprise and for which financing plan sequences can be collected are taken as enterprise samples. The enterprise samples are divided into multiple enterprise clusters according to the financial data sequence and credit sequence within the preset time period, and one enterprise cluster corresponds to one enterprise category; further, the average sequence of financing plans within an enterprise cluster is clustered in an ordered sample, and the preset time period is divided into multiple time sub-segments; the financing plans within each time sub-segment are basically the same, and the differences in financing plans between adjacent time sub-segments are large. When the variance of financial data in each time sub-segment is large, it means that changes in financial data will not cause changes in financing plans, and financial data have little impact on financing plans; between two adjacent time sub-segments, if the financial data between the two adjacent time sub-segments is The DTW distance is small, indicating that the financial data between two adjacent time sub-segments has not changed, that is, the change in the financing plan is not caused by the change in the financial data, and the impact of the financial data is small, and then the influence coefficient of the financial data and credit on the financing plan in the enterprise cluster can be calculated; in the enterprise cluster to which the target enterprise belongs, the second similarity between the target enterprise in the current financing cycle and the enterprise samples in the historical financing cycle is calculated based on the influence coefficient, and the financing plans of the enterprise samples in the historical financing cycle are weighted and summed according to the second similarity to obtain the recommended financing plan of the target enterprise, which can eliminate the financing plan errors caused by differences in business areas, financial conditions and information records between enterprises, and accurately obtain the financing plan of the target enterprise. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a flow chart of a data processing method based on multi-channel financing according to an embodiment of the present application. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0024] According to the first aspect of the present application, the present application provides a data processing method based on multi-channel financing, which is used to obtain a recommended financing plan for a target enterprise. Figure 1 is a flow chart of a data processing method based on multi-channel financing according to an embodiment of the present application. Figure 1 As shown, the data processing method based on multi-channel financing includes steps S101 to S104, which are described in detail below.

[0025] S101, obtaining financial data sequences, credit sequences and financing scheme sequences of multiple enterprise samples within a preset time period, wherein the preset time period includes multiple historical financing cycles, and the financing scheme is the proportion of financing volume of each financing channel within a financing cycle.

[0026] In one embodiment, the enterprise sample and the target enterprise belong to the same industry. For example, if the target enterprise is in the infrastructure industry, multiple enterprise samples are obtained in the infrastructure industry. The financing cycle can be one month, one quarter or one year, which can be set according to needs. A recommended financing plan will be generated for the target enterprise in each financing cycle.

[0027] It should be noted that not all enterprises can collect financing plan sequences. Therefore, when obtaining enterprise samples, enterprises that can collect financing plan sequences should be selected as enterprise samples.

[0028] For a corporate sample, the financial data, credit rating and financing scheme of multiple historical financing cycles are collected to obtain the financial data sequence, credit rating sequence and financing scheme sequence of the corporate sample, wherein the financial data includes at least one of liabilities, working capital, expenditure and income; the credit rating is the default rate within a financing cycle; the financing scheme is the proportion of financing volume of each financing channel within a financing cycle, wherein each financing channel corresponds to each financial institution. The number of financing channels is recorded as A, and within a financing cycle, the proportion of financing volume of each financing channel in the total financing volume is calculated to obtain the financing scheme of row A and column 1.

[0029] Among them, multiple historical financing cycles constitute a preset time period. Exemplarily, the number of historical financing cycles is set to 12, that is, the preset time period includes 12 historical financing cycles.

[0030] In this way, the financial data sequence, credit sequence and financing plan sequence of multiple enterprise samples within a preset time period are obtained; the financial data sequence and credit sequence can reflect the operating conditions and credit records of the enterprise samples, and are the basis for determining the financing plan.

[0031] S102, clustering enterprise samples according to the financial data sequence and the credit sequence to obtain multiple enterprise clusters.

[0032] In one embodiment, enterprise samples are clustered according to the financial data sequence and the credit sequence, and all enterprise samples are divided into multiple enterprise clusters, where one enterprise cluster includes multiple enterprise samples. Specifically, obtaining multiple enterprise clusters includes: clustering enterprise samples using a Kmeans clustering algorithm, determining the number of clusters using an elbow method, and obtaining multiple enterprise clusters, wherein the cluster distance between any enterprise samples in the Kmeans clustering algorithm is the sum of the DTW distance of the financial data sequence and the DTW distance of the credit sequence.

[0033] In this way, the enterprise samples are classified unsupervisedly and divided into multiple enterprise clusters. One enterprise cluster corresponds to one category, and the financial data and credit rating of each enterprise sample in a cluster are basically the same.

[0034] S103, perform ordered sample clustering on the average sequence of financing plans within the enterprise cluster, divide the preset time period into multiple time sub-segments, and calculate the influence coefficients of the financial data and credit in the enterprise cluster on the financing plan respectively; the influence coefficient of the financial data is negatively correlated with the variance of the financial data in the time sub-segment, and is positively correlated with the DTW distance of the financial data between adjacent time sub-segments.

[0035] In one embodiment, within any enterprise cluster, the financial data sequence and credit sequence of each enterprise sample are basically the same. In order to accurately obtain the impact of financial data and credit on the financing plan in the enterprise cluster, it is necessary to calculate the average sequence of financing plans of the enterprise cluster. The average sequence of financing plans includes the average financing plans in each historical financing cycle.

[0036] Furthermore, ordered sample clustering is performed on the average sequence of financing plans within the enterprise cluster, and the preset time period is divided into multiple time sub-segments. Specifically, the ordered sample clustering of the average sequence of financing plans within the enterprise cluster includes: performing ordered sample clustering on the average sequence of financing plans according to the initial number to obtain the clustering results of the initial number, and the clustering results include the initial number of initial time sub-segments; calculating the evaluation value of the initial number, and the evaluation value is negatively correlated with the variance of the average financing plan in the initial time sub-segment, and is positively correlated with the DTW distance of the adjacent initial time sub-segments; updating the initial number to draw the evaluation value curve, obtaining the clustering results of the initial number corresponding to the inflection point in the evaluation value curve, and obtaining multiple time sub-segments.

[0037] Among them, ordered sample clustering will generate multiple segmentation points for dividing the preset time period. When the number of segmentation points is 1, the preset time period is divided into 2 initial time sub-segments. As the number of segmentation points increases, the number of initial time sub-segments also increases; an initial time sub-segment includes multiple average financing plans. Ideally, the average financing plan in an initial time sub-segment should remain consistent, and the average financing plans in adjacent initial time sub-segments should differ greatly. Therefore, the smaller the variance of the average financing plan in the initial time sub-segment, the better the clustering effect, and the larger the corresponding evaluation value. In order to prevent the initial number from being too large and causing the time sub-segment to be too short, the initial number corresponding to the inflection point in the evaluation value curve is used as the target number, and the clustering results corresponding to the target number correspond to multiple time sub-segments.

[0038] In one embodiment, to avoid the influence of noise data on the influence coefficient, before respectively calculating the influence coefficients of the financial data and credit rating of the enterprise cluster on the financing plan, the processing method further includes: deleting time sub-segments whose length is less than a preset length.

[0039] It can be understood that the preset length is 3, that is, when the number of historical financing cycles in a time sub-segment is less than 3, the time sub-segment is considered to be noise data and is deleted to avoid the influence of noise data on "respectively calculating the influence coefficients of the financial data and credit rating of the enterprise cluster on the financing plan" and improve the accuracy of the influence coefficient.

[0040] In one embodiment, the impact coefficient of the financial data is negatively correlated with the variance of the financial data in the time sub-segment, and positively correlated with the DTW distance of the financial data between adjacent time sub-segments. It can be understood that the financing scheme in a time sub-segment remains unchanged. If the variance of the financial data in each time sub-segment is large, it means that the change of the financial data will not cause the change of the financing scheme, and the financial data has little impact on the financing scheme, that is, the impact coefficient of the financial data is small; between two adjacent time sub-segments, the financing scheme has changed. If the DTW distance of the financial data between the two adjacent time sub-segments is small, it means that the financial data between the two adjacent time sub-segments has not changed, that is, the change of the financing scheme is not caused by the change of the financial data, that is, the impact coefficient of the financial data is small.

[0041] In summary, enterprise clusters The impact coefficient of financial data for: , For time sub-segment and time sub-segments The DTW distance of financial data, is the number of time sub-segments, For time sub-segment The variance of financial data.

[0042] In the same way, the enterprise clusters can be obtained based on the variance of credit in the time sub-segment and the DTW distance of credit between time sub-segments. The influence coefficient of medium credit , I will not go into details here.

[0043] In this way, taking into account the large differences in financial data and creditworthiness among different enterprise clusters, and avoiding the impact of the changing trend of financial data and creditworthiness within a preset time period on the impact coefficient, the impact coefficient of financial data and creditworthiness on the financing plan is calculated in each enterprise cluster to ensure the accuracy of the impact coefficient.

[0044] S104, in the enterprise cluster to which the target enterprise belongs, the second similarity between the target enterprise in the current financing cycle and each enterprise sample in the historical financing cycle is calculated based on the influence coefficient of financial data and creditworthiness, and the financing plans of each enterprise sample in the historical financing cycle are weighted and summed based on the second similarity to obtain a recommended financing plan for the target enterprise.

[0045] In one embodiment, a method for obtaining the enterprise cluster to which a target enterprise belongs includes: collecting the financial data sequence and credit sequence of the target enterprise within a preset time period before the current financing cycle; calculating the first similarity between the target enterprise and each enterprise cluster, and taking the enterprise cluster corresponding to the maximum sequence similarity as the enterprise cluster to which the target enterprise belongs; the first similarity is the sequence similarity between the financial data sequence and the average sequence of financial data in the enterprise cluster, and the average of the sequence similarity between the credit sequence and the average sequence of credit in the enterprise cluster.

[0046] Among them, in order to obtain the recommended financing plan for the target enterprise's current financing cycle, the financial data sequence and credit sequence of the target enterprise in the preset time period before the current financing cycle are collected. The financial data sequence and credit sequence can reflect the operating conditions and credit record of the target enterprise when obtaining the recommended financing plan.

[0047] After determining the enterprise cluster to which the target enterprise belongs, the enterprise cluster to which the target enterprise belongs includes the financial data, credit rating and financing plans of enterprise samples in multiple historical financing cycles. For example, the historical financing cycle Financial data, creditworthiness and financing plans of sample c of Chinese enterprises.

[0048] The second similarity between the target enterprise in the current financing cycle and each enterprise sample in the historical financing cycle is calculated based on the influence coefficient of financial data and credit. The second similarity of enterprise sample c for: ; and is the influence coefficient of financial data and creditworthiness in the enterprise cluster to which the target enterprise belongs, and The target companies and historical financing cycles are The distance similarity of financial data and creditworthiness in the internal enterprise sample c.

[0049] Among them, the distance similarity of financial data is used to reflect the target enterprise in the current financing cycle and the historical financing cycle The similarity of the financial data of enterprise sample c in , the distance similarity of financial data is negatively correlated with the absolute value of the difference between the financial data. When the absolute value of the difference between the financial data is 0, the distance similarity of financial data is equal to 1.

[0050] In one embodiment, after obtaining the second similarity between the target enterprise and each enterprise sample in the historical financing cycle, the financing schemes of each enterprise sample in the historical financing cycle are weighted and summed according to the second similarity to obtain the recommended financing scheme of the target enterprise in the current financing cycle. Specifically, obtaining the recommended financing scheme of the target enterprise includes: using the normalized second similarity to weightedly sum the financing schemes of each enterprise sample in the historical financing cycle to obtain the recommended financing scheme of the target enterprise.

[0051] In another embodiment, in order to avoid the influence of the financing scheme whose second similarity is less than the similarity threshold on the recommended financing scheme, the recommended financing scheme of the target enterprise is calculated based on the financing scheme corresponding to the enterprise samples in the historical financing cycle whose second similarity is greater than the similarity threshold. Specifically, obtaining the recommended financing scheme of the target enterprise includes: after deleting the enterprise samples in multiple historical financing cycles whose second similarity is not greater than the similarity threshold, weighted summing the financing schemes of each enterprise sample in the historical financing cycle by using the normalized second similarity, and obtaining the recommended financing scheme of the target enterprise. The similarity threshold is 0.5.

[0052] In this way, the enterprise cluster to which the target enterprise belongs includes the financial data, creditworthiness and financing plans of each enterprise sample in the historical financing cycle; the second similarity between the target enterprise and each enterprise sample in the historical financing cycle is calculated based on the financial data and creditworthiness, and then the financing plans of each enterprise sample in the historical financing cycle are weighted and summed according to the second similarity to obtain the recommended financing plan for the target enterprise in the current financing cycle, thereby eliminating the financing plan errors caused by differences in business areas, financial conditions and information records between enterprises, and accurately obtaining the financing plan of the target enterprise.

[0053] The above describes the technical principle and implementation details of a data processing method based on multi-channel financing of this application through specific embodiments. By comparing the financial data sequence and credit sequence of the target enterprise in the current financing cycle with the enterprise samples in the historical financing cycle, the recommended financing plan of the target enterprise is obtained, which can eliminate the financing plan errors caused by differences in business areas, financial conditions and information records between enterprises, and accurately obtain the financing plan of the target enterprise.

[0054] It should be noted that, for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present application, and these modifications and improvements are all within the scope of protection of the present application. Therefore, the scope of protection of the patent of the present application shall be subject to the attached claims.

Claims

1. A data processing method based on multi-channel financing, characterized in that: The processing method includes: obtaining a financial data sequence, a credit sequence, and a financing plan sequence of multiple enterprise samples within a preset time period, wherein the preset time period includes multiple historical financing cycles, and the financing plan is the proportion of the financing amount of each financing channel within a financing cycle; Clustering enterprise samples based on financial data series and credit series to obtain multiple enterprise clusters; Orderly sample clustering is performed on the average sequence of financing plans within the enterprise cluster, and the preset time period is divided into multiple time sub-segments, and the influence coefficients of the financial data and credit rating in the enterprise cluster on the financing plan are calculated respectively; the influence coefficient of financial data is negatively correlated with the variance of financial data in the time sub-segment, and positively correlated with the DTW distance of financial data between adjacent time sub-segments; In the enterprise cluster to which the target enterprise belongs, the second similarity between the target enterprise in the current financing cycle and the enterprise samples in the historical financing cycle is calculated based on the influence coefficient of financial data and creditworthiness. The financing plans of the enterprise samples in the historical financing cycle are weighted and summed according to the second similarity to obtain the recommended financing plan for the target enterprise.

2. A data processing method based on multi-channel financing according to claim 1, characterized in that: The financial data includes at least one of liabilities, current funds, expenditures and income; the credit rating is the default rate within a financing cycle.

3. A data processing method based on multi-channel financing according to claim 1, characterized in that: Get multiple enterprise clusters including: The enterprise samples are clustered using the Kmeans clustering algorithm, and the number of clusters is determined using the elbow method to obtain multiple enterprise clusters. The clustering distance between any enterprise samples in the Kmeans clustering algorithm is the sum of the DTW distance of the financial data sequence and the DTW distance of the credit sequence.

4. A data processing method based on multi-channel financing according to claim 1, characterized in that: Ordered sample clustering of the average sequence of financing plans within the enterprise cluster includes: Performing ordered sample clustering on the average sequence of the financing schemes according to the initial number to obtain the clustering results of the initial number, wherein the clustering results include the initial number of initial time sub-segments; Calculating an evaluation value of the initial number, wherein the evaluation value is negatively correlated with the variance of the average financing scheme in the initial time sub-segment and positively correlated with the DTW distance of adjacent initial time sub-segments; The initial number is updated to draw the evaluation value curve, and the clustering results of the initial number corresponding to the inflection points in the evaluation value curve are obtained to obtain a plurality of time sub-segments.

5. The data processing method based on multi-channel financing according to claim 1 is characterized in that: Before respectively calculating the influence coefficients of the financial data and credit ratings in the enterprise cluster on the financing plan, the processing method further includes: deleting time sub-segments whose length is less than a preset length.

6. The data processing method based on multi-channel financing according to claim 1 is characterized in that: Enterprise Cluster The impact coefficient of financial data for: , For time sub-segment and time sub-segments The DTW distance of financial data, is the number of time sub-segments, For time sub-segment The variance of financial data.

7. The data processing method based on multi-channel financing according to claim 1 is characterized in that: Methods for obtaining the enterprise cluster to which the target enterprise belongs include: Collect the target enterprise's financial data series and credit series within a preset time period before the current financing cycle; Calculate the first similarity between the target enterprise and each enterprise cluster, and take the enterprise cluster corresponding to the maximum sequence similarity as the enterprise cluster to which the target enterprise belongs; the first similarity is the average of the sequence similarity between the financial data sequence and the average sequence of financial data in the enterprise cluster, and the sequence similarity between the credit sequence and the average sequence of credit in the enterprise cluster.

8. The data processing method based on multi-channel financing according to claim 1 is characterized in that: Target companies and historical financing cycles The second similarity of enterprise sample c for: ; and is the influence coefficient of financial data and creditworthiness in the enterprise cluster to which the target enterprise belongs, and The target companies and historical financing cycles are The distance similarity of financial data and creditworthiness in the internal enterprise sample c.

9. The data processing method based on multi-channel financing according to claim 1 is characterized in that: The recommended financing solutions for the target enterprises include: The normalized second similarity is used to perform weighted summation on the financing plans of each enterprise sample in the historical financing cycle to obtain the recommended financing plan for the target enterprise.

10. A data processing method based on multi-channel financing according to claim 9, characterized in that: Before obtaining the recommended financing plan for the target enterprise, the processing method further includes: deleting enterprise samples within multiple historical financing cycles whose second similarity is not greater than a similarity threshold.

Images