Financial management consultation method and system based on big data

Abstract

The invention relates to the technical field of financial management, in particular to a financial management consultation method and system based on big data. Obtaining a fund raising use plan of the current consultation object, extracting reference features of the real-time financial data based on the big data and the fund raising use plan, and obtaining a threatening key index through a clustering result of a plurality of key indexes obtained through the reference features and the big data; and constructing a financial mutation trend prediction model by calculating a probability quality function and a time interval distribution rule of mutation of historical financial derivative data, and predicting real-time data by using the financial mutation trend prediction model to obtain a future financial mutation trend of the threatening key index. Quality and efficiency of financial management consultation are improved, and reliability is high.

Description

Technical Field

[0001] The present invention relates to the field of financial management technology, and in particular to a financial management consulting method and system based on big data. Background Art

[0002] Financial management is the use of management methods to plan, organize and control the raising, allocation and use of funds to achieve the financial goals of enterprises or organizations. Its main goal is to maximize the value of enterprises or wealth through effective fund management. Before financial management, users usually conduct some corresponding accurate and low-risk consultations and strategies to increase financial management returns or minimize losses; therefore, financial management consultation, as a permanent business of some institutions, can further facilitate the financial management life of the masses and enhance the financial management activity of the masses. However, the existing financial management consultation methods still have some drawbacks. In the consultation process, large errors may occur through simple threshold comparison, and it is impossible to combine and analyze with real-time updated big financial data, and usually requires manual intervention to correct errors, which greatly reduces the efficiency of consultation and the accuracy of financial management consultation is low; at the same time, it is difficult to analyze the feasibility of the financial management direction consulted by the current user by predicting the trend of some key indicators that affect financial management behavior, so it is difficult to further conduct accurate risk warnings and formulate reasonable management optimization plans. Summary of the invention

[0003] The present invention overcomes the deficiencies of the prior art and provides a financial management consulting method and system based on big data.

[0004] To achieve the above object, the technical solution adopted by the present invention is: The first aspect of the present invention provides a financial management consulting method based on big data, comprising the following steps: S102: Obtain the fund-raising use plan of the current consulting object, extract the benchmark features of the real-time financial data based on the big data and the fund-raising use plan, and obtain the threatening key indicators through the clustering results of the benchmark features and several key indicators obtained from the big data; S104: Obtain historical financial derivative data generated by threatening key indicators, calculate the probability mass function of mutations in the historical financial derivative data and the time interval distribution law to construct a financial mutation trend prediction model, and use the financial mutation trend prediction model to predict real-time data to obtain future financial mutation trends of threatening key indicators; S106: Based on the big data, a number of historical financial loss data under the condition of the future sudden financial trend are obtained, and the impact of the future sudden financial trend of the threatening key indicator on the corresponding financial management purpose is calculated by regression based on the historical financial loss data, and the loss risk status of the amount of funds raised for financial management purposes is analyzed based on the impact range to obtain financial management consulting early warning information; S108: Eliminate and optimize the loss amount of the current consulting object when applying the raised funds for financial management purposes through the loss deviation of abnormal losses in the financial management consulting early warning information, and plan the allocation of the raised funds in the ideal financial management model to generate a financial management consulting optimization plan.

[0005] More specifically, the step S102 includes the following steps: Obtain the fund raising and use plan of the current consulting object, and extract the financial management purpose of the fund raising and use plan of the current consulting object; wherein the financial management purpose includes investment projects, daily operations and strategic expenditure activities; Based on big data retrieval, historical financial data and real-time financial data related to the financial management purpose are obtained, and the statistical values ​​of changes in the historical financial data are counted by constructing an autonomous encoder and an autonomous decoder to train the sliding window to reconstruct the features expressing the real-time financial data, thereby obtaining the benchmark features of the real-time financial data; Using big data to obtain several key indicators that lead to or affect the benchmark characteristics, randomly extracting a portion of data samples of the same size from each group of real-time financial data, defining the data as a sub-sample data set, and obtaining multiple sub-sample data sets; An agglomerative hierarchical clustering algorithm is introduced, each data in the sub-sample data set is defined as a cluster point, and the Mahalanobis distance matrix between adjacent cluster points is obtained. Based on the Mahalanobis distance matrix, each of the multiple sub-sample data sets is merged and clustered in the agglomerative hierarchical clustering algorithm to obtain a series of new cluster points. According to a series of new cluster points, the corresponding cluster points are returned for iterative update, and the above cluster point merging clustering and iterative update steps are repeated continuously until all data points are clustered, and an agglomerative hierarchical clustering tree is generated; The leaf node values ​​output by the final hierarchical clustering are obtained through the agglomerative hierarchical clustering tree, and one or more key indicators that affect the current consulting object's performance of financial management purposes are determined based on the leaf node values, and are defined as threatening key indicators.

[0006] More specifically, the step S102 of acquiring historical financial data and real-time financial data related to the financial management purpose based on big data retrieval, and constructing an autonomous encoder and an autonomous decoder to train a sliding window to count the change statistics of the historical financial data to reconstruct the features expressing the real-time financial data, and obtain the benchmark features of the real-time financial data, specifically includes the following steps: Preset the search frequency and established statistical indicators of financial data in big data, and retrieve multiple groups of historical financial data related to financial management purposes within a preset time period in big data according to the search frequency; Setting a sliding window specification based on the search frequency, constructing a sliding window in the plurality of groups of the real-time financial data according to the sliding window specification, and counting the change amount of a predetermined statistical indicator of the data sliding window at an initial position of the historical financial data; After the statistics are completed, the data sliding window is moved to enter the next position in the historical financial data to count the change amount of the predetermined statistical indicator, and the moving data sliding window is repeatedly moved to count the change amount until all data points of the historical financial data are traversed, and finally the historical change statistical value of the predetermined statistical indicator is output; Construct an autonomous encoder model and an autonomous decoder model of a neural network layer architecture, train the historical change statistics of the established statistical indicators through the autonomous encoder model, and generate a mean square error function; Introducing a back propagation algorithm to calculate the mean square error function, obtaining a gradient value of the mean square error function compared to the bias and weight in the autonomous encoder model, iteratively updating the neural network parameters of the autonomous encoder model based on the gradient value, and obtaining a trained autonomous encoder model; Based on big data, multiple groups of real-time financial data for related financial management purposes are obtained, and the trained autonomous encoder model is used to operate on the multiple groups of real-time financial data. The autonomous decoder model is used to decode the low-dimensional financial data features in the autonomous encoder model and reconstruct the expression to obtain the baseline features of the real-time financial data.

[0007] More specifically, the step S104 includes the following steps: Preset a predetermined time interval, obtain N historical financial derivative data generated when the threatening key indicator is within the predetermined time interval, perform mutation analysis on the N historical financial derivative data through the benchmark characteristics of the real-time financial data, and obtain the mutation frequency of the historical financial derivative data; Constructing a Poisson distribution model, estimating a strength threshold of data mutation according to the mutation frequency, performing random variable distribution calculation in the Poisson distribution model based on the strength threshold, and obtaining a probability mass function of mutation of historical financial derivative data; Analyze the global waiting time of sudden changes in historical financial derivative data according to the probability mass function, and construct the global waiting time by using the exponential distribution method to obtain the time interval distribution law of sudden changes in historical financial derivative data; Based on the probability mass function and the time interval distribution law, the evolution trend of the sudden change of historical financial derivative data in the Poisson distribution process is cumulatively described, and the variance value of the probability mass function in the time interval distribution law is obtained during the cumulative evolution trend; Check whether the variance value is greater than or equal to the preset variance value. If so, continue to accumulate the description; if less than, recalculate and accumulate the probability mass function and the time interval distribution law until the variance value is greater than or equal to the preset variance value, and generate a financial mutation trend prediction model; N real-time financial derivative data generated by threatening key indicators within a predetermined time interval are obtained, and the real-time financial derivative data are predicted by a financial mutation trend prediction model to obtain the future financial mutation trend of the threatening key indicators.

[0008] More specifically, the step S106 includes the following steps: Based on big data, a number of historical financial loss data for financial management purposes are obtained within a preset time period under the condition of future financial sudden changes in the threatening key indicators; Calculate the correlation coefficient between the financial management purpose and the future financial sudden change trend of the threatening key indicators based on the historical financial loss data, and establish a correlation coefficient matrix; Constructing a regression equation based on the correlation coefficient matrix, substituting the historical financial loss data into the regression equation to describe the impact relationship between future financial sudden changes and financial management purposes, and generating an impact relationship regression equation; Solving the impact relationship regression equation to obtain a plurality of partial regression coefficients, and determining the impact of the future financial mutation trend of the threatening key indicator on the corresponding financial management purpose according to the partial regression coefficients; Obtaining the intended consulting information on financial management submitted by the current consulting object, extracting the critical loss coefficient that the current consulting object can tolerate through the intended consulting information, and evaluating the critical loss coefficient based on the financial management risk standard to set the risk loss amount range of the current consulting object; Obtain the amount of funds raised by the current consulting object for financial management purposes, and estimate the estimated amount of loss that would be incurred after the amount of funds raised is used for financial management purposes based on the impact range; If the estimated loss amount is not within the risk loss amount range, the estimated loss amount is calibrated as a normal loss; if it is within the risk loss amount range, the estimated loss amount is calibrated as an abnormal loss, and a risk warning signal and risk warning level are generated based on the abnormal loss and uploaded to the financial management consulting terminal to obtain financial management consulting warning information.

[0009] More specifically, the step S108 includes the following steps: Obtaining the lower limit loss amount of the risk loss amount range, calculating the difference between the estimated loss amount of the abnormal loss of the financial management consulting early warning information and the lower limit loss amount, and obtaining the amount deviation value; Obtaining the ideal financial management model of the current consulting object for applying the raised funds amount to financial management purposes through the intended consulting information; wherein the financial management model includes asset allocation, investment portfolio and cost control; Based on the amount deviation value, the application of the loss gradient of the raised funds amount when generating abnormal losses for financial management purposes is gradually optimized until the amount deviation value is eliminated, thereby obtaining the loss optimization parameter; Based on the loss optimization parameters, the allocation of the amount of funds raised in the ideal financial management model is calculated and planned to generate a financial management allocation network, and the financial management allocation network is deployed on the actual financial management model to optimize the use of the amount of funds raised and generate a financial management consulting optimization plan.

[0010] The second aspect of the present invention provides a financial management consulting system based on big data, the financial management consulting system includes a memory and a processor, the memory stores a financial management consulting method program based on big data, when the financial management consulting method program is executed by the processor, any one of the steps of the financial management consulting method is implemented.

[0011] The present invention solves the technical defects existing in the background technology, and the beneficial technical effects of the present invention are: Obtain the fund-raising use plan of the current consulting object, extract the benchmark features of real-time financial data based on big data and the fund-raising use plan, and obtain threatening key indicators through clustering results of several key indicators obtained through benchmark features and big data; construct a financial mutation trend prediction model by calculating the probability mass function and time interval distribution law of mutations in historical financial derivative data, and use the financial mutation trend prediction model to predict real-time data to obtain future financial mutation trends of threatening key indicators; obtain several historical financial loss data under the conditions of the future financial mutation trend based on big data, and calculate the impact of the future financial mutation trend of threatening key indicators on the corresponding financial management purposes through regression based on the historical financial loss data, analyze the loss risk status of the amount of funds raised for financial management purposes based on the impact, and obtain financial management consulting early warning information; eliminate and optimize the loss amount of the current consulting object when the amount of funds raised is applied to financial management purposes through the loss deviation of abnormal losses in the financial management consulting early warning information, and plan the allocation of the amount of funds raised in the ideal financial management model to generate a financial management consulting optimization plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, drawings of other embodiments can be obtained based on these drawings without paying creative work.

[0013] Figure 1 A first method flow chart of a financial management consulting method based on big data is shown; Figure 2 A second method flow chart of a financial management consulting method based on big data is shown; Figure 3 A third method flow chart of a financial management consulting method based on big data is shown; Figure 4 A system framework diagram of a financial management consulting system based on big data is shown. DETAILED DESCRIPTION

[0014] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0015] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0016] The first aspect of the present invention provides a financial management consulting method based on big data, such as Figure 1 As shown, the following steps are included: S102: Obtain the fund-raising use plan of the current consulting object, extract the benchmark features of the real-time financial data based on the big data and the fund-raising use plan, and obtain the threatening key indicators through the clustering results of the benchmark features and several key indicators obtained from the big data; S104: Obtain historical financial derivative data generated by threatening key indicators, calculate the probability mass function of mutations in the historical financial derivative data and the time interval distribution law to construct a financial mutation trend prediction model, and use the financial mutation trend prediction model to predict real-time data to obtain future financial mutation trends of threatening key indicators; S106: Based on the big data, a number of historical financial loss data under the condition of the future sudden financial trend are obtained, and the impact of the future sudden financial trend of the threatening key indicator on the corresponding financial management purpose is calculated by regression based on the historical financial loss data, and the loss risk status of the amount of funds raised for financial management purposes is analyzed based on the impact range to obtain financial management consulting early warning information; S108: Eliminate and optimize the loss amount of the current consulting object when applying the raised funds for financial management purposes through the loss deviation of abnormal losses in the financial management consulting early warning information, and plan the allocation of the raised funds in the ideal financial management model to generate a financial management consulting optimization plan.

[0017] More specifically, the step S102 is as follows: Figure 2 As shown, the specific steps include: S202: Obtain the fund raising and use plan of the current consulting object, and extract the financial management purpose of the fund raising and use plan of the current consulting object; wherein the financial management purpose includes investment projects, daily operations and strategic expenditure activities; S204: Acquire historical financial data and real-time financial data related to the financial management purpose based on big data retrieval, and construct an autonomous encoder and an autonomous decoder to train a sliding window to count the change statistics of the historical financial data, so as to reconstruct the features expressing the real-time financial data and obtain the benchmark features of the real-time financial data; S206: using big data to obtain several key indicators that lead to the generation of the benchmark characteristics or affect the benchmark characteristics, randomly extracting a portion of data samples of the same size from each group of the real-time financial data to define the data as a sub-sample data set, and obtaining multiple sub-sample data sets; S208: introducing an agglomerative hierarchical clustering algorithm, defining each data in the sub-sample data set as a cluster point, obtaining a Mahalanobis distance matrix between adjacent cluster points, and merging and clustering each of the multiple sub-sample data sets in the agglomerative hierarchical clustering algorithm based on the Mahalanobis distance matrix to obtain a series of new cluster points; S210: returning the corresponding clustering points for iterative updating according to a series of new clustering points, and continuously repeating the above clustering point merging and clustering and iterative updating steps until all data points are clustered, thereby generating an agglomerative hierarchical clustering tree; S212: Obtain the leaf node value output by the final hierarchical clustering through the agglomerative hierarchical clustering tree, and determine one or more key indicators that affect the current consulting object's performance of financial management purposes based on the leaf node value, and define them as threatening key indicators.

[0018] It should be noted that the benchmark characteristics of real-time financial data include high frequency, volatility, aggregation, nonlinearity, time series, non-stationarity and uncertainty. The key indicators include GDP (gross domestic product), income and profit, interest rate, exchange rate, inflation, price-earnings ratio, price-to-book ratio and debt-to-asset ratio. For some specific financial consulting users, the funds raised are usually used for investment projects, daily operations and strategic expenditure activities. Therefore, these users will consult certain key indicators that have a threatening impact on their financial management purposes through banks and other financial institutions. For example, financial investment is usually affected by key indicators such as economic cycles, inflation or exchange rates, and the changes in these key indicators will directly or indirectly radiate their financial investment activities. However, existing financial management consulting is usually difficult to accurately locate threatening key indicators, or the located indicators have errors, so that the consulting objects are greatly affected by the changes in these threatening key indicators when managing different fund purposes, causing their financial management to have large losses and management loopholes, greatly reducing the efficiency and accuracy of financial management. Therefore, this method can accurately cluster the key indicators that may pose a threat to the consultant when the financial management purpose is executed based on the baseline characteristics of the financial management purpose consulted by the consultant. With the support of big data technology, the consultant can pay attention to and attach importance to these threatening key indicators, understand the impact of these threatening indicators on the required financial management purposes, and improve the accuracy of subsequent financial management risk analysis and avoidance.

[0019] More specifically, in step S102, historical financial data and real-time financial data related to the financial management purpose are retrieved based on big data, and the statistical values ​​of changes in historical financial data are counted by constructing an autonomous encoder and an autonomous decoder to train a sliding window to reconstruct the features expressing the real-time financial data, thereby obtaining the benchmark features of the real-time financial data, such as Figure 3 As shown, the specific steps include: S302: Preset the search frequency of financial data in big data and the established statistical indicators, and retrieve multiple groups of historical financial data related to financial management purposes within a preset time period in the big data according to the search frequency; S304: setting a sliding window specification based on the search frequency, constructing a sliding window in the plurality of groups of the real-time financial data according to the sliding window specification, and counting the change amount of a predetermined statistical indicator of the data sliding window at an initial position of the historical financial data; S306: after the statistics are completed, the data sliding window is moved to enter the next position in the historical financial data to count the change amount of the predetermined statistical indicator, and the moving data sliding window is repeatedly moved to count the change amount until all data points of the historical financial data are traversed, and finally the historical change statistical value of the predetermined statistical indicator is output; S308: constructing an autonomous encoder model and an autonomous decoder model of a neural network layer architecture, training historical change statistics of a given statistical indicator through the autonomous encoder model, and generating a mean square error function; S310: Introducing a back propagation algorithm to calculate the mean square error function, obtaining a gradient value of the mean square error function compared to the bias and weight in the autonomous encoder model, iteratively updating the neural network parameters of the autonomous encoder model based on the gradient value, and obtaining a trained autonomous encoder model; S312: Based on big data, multiple groups of real-time financial data for related financial management purposes are obtained, the multiple groups of real-time financial data are operated through the trained autonomous encoder model, and the low-dimensional financial data features in the autonomous encoder model are decoded and reconstructed using the autonomous decoder model to obtain the baseline features of the real-time financial data.

[0020] It should be noted that financial data includes transaction data, market data, economic data, social media information, news, and financial statements. Some key indicators of financial management data have certain variability, such as sharp changes in exchange rate indicators or financial asset prices or yields, which have certain volatility and aggregation; these amplitude fluctuations may be high-frequency, so there is high frequency; and its volatility, aggregation or high frequency are the basic characteristics of financial data. Therefore, in order to accurately cluster and obtain threatening key indicators that affect financial management purposes, it is necessary to first extract the basic features in its financial data. Through this method, it is possible to describe the changing trend of its financial data in a sliding window moving manner and reconstruct its changing trend through the autonomous encoder model and the autonomous decoder model, so that compared with the existing consulting methods, it can more accurately extract all the benchmark features existing in the financial data, and then effectively improve the accuracy of the subsequent real-time financial data clustering key indicators, and ensure the reliability of the positioning of threatening key indicators.

[0021] More specifically, the step S104 includes the following steps: Preset a predetermined time interval, obtain N historical financial derivative data generated when the threatening key indicator is within the predetermined time interval, perform mutation analysis on the N historical financial derivative data through the benchmark characteristics of the real-time financial data, and obtain the mutation frequency of the historical financial derivative data; Constructing a Poisson distribution model, estimating a strength threshold of data mutation according to the mutation frequency, performing random variable distribution calculation in the Poisson distribution model based on the strength threshold, and obtaining a probability mass function of mutation of historical financial derivative data; Analyze the global waiting time of sudden changes in historical financial derivative data according to the probability mass function, and construct the global waiting time by using the exponential distribution method to obtain the time interval distribution law of sudden changes in historical financial derivative data; Based on the probability mass function and the time interval distribution law, the evolution trend of the sudden change of historical financial derivative data in the Poisson distribution process is cumulatively described, and the variance value of the probability mass function in the time interval distribution law is obtained during the cumulative evolution trend; Check whether the variance value is greater than or equal to the preset variance value. If so, continue to accumulate the description; if less than, recalculate and accumulate the probability mass function and the time interval distribution law until the variance value is greater than or equal to the preset variance value, and generate a financial mutation trend prediction model; N real-time financial derivative data generated by threatening key indicators within a predetermined time interval are obtained, and the real-time financial derivative data are predicted by a financial mutation trend prediction model to obtain the future financial mutation trend of the threatening key indicators.

[0022] It should be noted that after obtaining the threatening key indicators in the process of the consulting object's financial management purposes, since the changes or sudden changes of these threatening key indicators will directly or indirectly affect the benefits of raising funds in the future financial management purposes, it is necessary to predict the trend of the sudden changes of these threatening key indicators, so as to deeply analyze whether it is feasible to maximize the benefits of raising funds for financial management purposes in the future. Therefore, this method uses the Poisson distribution model to reveal the probability mass function and time interval distribution law of the sudden changes of threatening key indicators based on the mutation frequency of historical financial derivative data, so as to construct a prediction model that can predict its financial mutation trend. This model can accurately and efficiently predict the future financial mutation trend of threatening key indicators, such as the trend of a continuous decline in exchange rates in the future time period, etc.; It provides a reliable financial management analysis basis for consulting objects to use the rationalization of raising funds and maximize the benefits to perform financial management purposes, so as to reduce the risks brought by threatening key indicators to the use of consulting objects' funds.

[0023] More specifically, the step S106 includes the following steps: Based on big data, a number of historical financial loss data for financial management purposes are obtained within a preset time period under the condition of future financial sudden changes in the threatening key indicators; Calculate the correlation coefficient between the financial management purpose and the future financial sudden change trend of the threatening key indicators based on the historical financial loss data, and establish a correlation coefficient matrix; Constructing a regression equation based on the correlation coefficient matrix, substituting the historical financial loss data into the regression equation to describe the impact relationship between future financial sudden changes and financial management purposes, and generating an impact relationship regression equation; Solving the impact relationship regression equation to obtain a plurality of partial regression coefficients, and determining the impact of the future financial mutation trend of the threatening key indicator on the corresponding financial management purpose according to the partial regression coefficients; Obtaining the intended consulting information on financial management submitted by the current consulting object, extracting the critical loss coefficient that the current consulting object can tolerate through the intended consulting information, and evaluating the critical loss coefficient based on the financial management risk standard to set the risk loss amount range of the current consulting object; Obtain the amount of funds raised by the current consulting object for financial management purposes, and estimate the estimated amount of loss that would be incurred after the amount of funds raised is used for financial management purposes based on the impact range; If the estimated loss amount is not within the risk loss amount range, the estimated loss amount is calibrated as a normal loss; if it is within the risk loss amount range, the estimated loss amount is calibrated as an abnormal loss, and a risk warning signal and risk warning level are generated based on the abnormal loss and uploaded to the financial management consulting terminal to obtain financial management consulting warning information.

[0024] It should be noted that when the future financial trends of threatening key indicators change suddenly, they will inevitably have a certain degree of impact on the consulting object's financial management purposes. For example, inflation may cause losses in the consulting object's investment activities, causing the consulting object's investment returns to decline. Therefore, this method can accurately calculate and quantify the impact of threatening key indicators on the consulting object's financial management purposes, so that users can make corresponding response decisions in a timely manner during the consultation process to avoid changes in their threatening key indicators, thereby ensuring that the expected benefits are achieved when financial management purposes are implemented. When the impact range is detected, if the existing financial management consulting method is used at this time, there may be an indiscriminate warning phenomenon, that is, regardless of whether the evaluation result of the impact range is large or small, a warning signal will be issued. At this time, it is easy to cause panic among the consulting objects about financial management, and the psychological impact may cause a series of chain reactions such as errors in judgment in subsequent financial management operations, resulting in loss of profits; therefore, this method can be based on the impact range of future financial mutation trends of threatening key indicators on the execution of financial management purposes. In the process of user financial management consulting, the impact range will cause the loss of the amount of funds raised to execute financial management purposes and analyze it. Among them, if the estimated loss amount is not within the risk loss amount range, it means that the loss amount of the raised funds after executing financial management purposes under its impact range is small and within the expected acceptable range, so the estimated loss amount is calibrated as a normal loss; otherwise, it means that the loss amount is small and has exceeded the expected allowable loss amount range, and there is a high-risk financial management impact, so the estimated loss amount is calibrated as an abnormal loss. At this time, the high risk of the impact range is warned through early warning signals. Compared with the simple comparative analysis by setting the impact range threshold in the existing financial management consulting method, it can make the financial management risk prevention and avoidance more reasonable, reduce the frequency of manual intervention, save consulting time, reduce management losses and improve returns.

[0025] More specifically, the step S108 includes the following steps: Obtaining the lower limit loss amount of the risk loss amount range, calculating the difference between the estimated loss amount of the abnormal loss of the financial management consulting early warning information and the lower limit loss amount, and obtaining the amount deviation value; Obtaining the ideal financial management model of the current consulting object for applying the raised funds amount to financial management purposes through the intended consulting information; wherein the financial management model includes asset allocation, investment portfolio and cost control; Based on the amount deviation value, the application of the loss gradient of the raised funds amount when generating abnormal losses for financial management purposes is gradually optimized until the amount deviation value is eliminated, thereby obtaining the loss optimization parameter; Based on the loss optimization parameters, the allocation of the amount of funds raised in the ideal financial management model is calculated and planned to generate a financial management allocation network, and the financial management allocation network is deployed on the actual financial management model to optimize the use of the amount of funds raised and generate a financial management consulting optimization plan.

[0026] It should be noted that when the impact of the future trend of the threatening key indicators will cause the consulting object to suffer losses in the execution of financial management purposes, the loss effect should be eliminated or reduced to the maximum extent for the consulting user, ensuring that the benefits of the raised funds in the financial management purposes are maximized. Therefore, this method can optimize the losses caused by the impact of the impact, eliminate the loss phenomenon when the raised funds are used for financial management purposes, and make decisions for the consulting object to make the best financial management plan without losses; after eliminating the losses, the consulting user can reasonably plan the allocation of the raised funds in the ideal financial management model, such as the optimization of the investment portfolio, etc.; thereby improving the rationality of the management optimization of the raised funds by the consulting user and improving the economic benefits.

[0027] The second aspect of the present invention provides a financial management consulting system based on big data, such as Figure 4 As shown, the financial management consulting system includes a memory 41 and a processor 42. The memory 41 stores a financial management consulting method program based on big data. When the financial management consulting method program is executed by the processor 42, any one of the steps of the financial management consulting method is implemented.

[0028] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A financial management consulting method based on big data, characterized in that: The following steps are involved: S102: Obtain the fund-raising use plan of the current consulting object, extract the benchmark features of the real-time financial data based on the big data and the fund-raising use plan, and obtain the threatening key indicators through the clustering results of the benchmark features and several key indicators obtained from the big data; S104: Obtain historical financial derivative data generated by threatening key indicators, calculate the probability mass function of mutations in the historical financial derivative data and the time interval distribution law to construct a financial mutation trend prediction model, use the financial mutation trend prediction model to predict real-time data, and obtain the future financial mutation trend of threatening key indicators; S106: Based on the big data, a number of historical financial loss data under the condition of the future sudden financial trend are obtained, and the impact of the future sudden financial trend of the threatening key indicator on the corresponding financial management purpose is calculated by regression based on the historical financial loss data, and the loss risk status of the amount of funds raised for financial management purposes is analyzed based on the impact range to obtain financial management consulting early warning information; S108: Eliminate and optimize the loss amount of the current consulting object when applying the raised funds for financial management purposes through the loss deviation of abnormal losses in the financial management consulting early warning information, and plan the allocation of the raised funds in the ideal financial management model to generate a financial management consulting optimization plan.

2. A financial management consulting method based on big data according to claim 1, characterized in that: The step S102 specifically includes the following steps: Obtain the fund raising and use plan of the current consulting object, and extract the financial management purpose of the fund raising and use plan of the current consulting object; wherein the financial management purpose includes investment projects, daily operations and strategic expenditure activities; Based on big data retrieval, historical financial data and real-time financial data related to the financial management purpose are obtained, and the statistical values ​​of changes in the historical financial data are counted by constructing an autonomous encoder and an autonomous decoder to train a sliding window to reconstruct the features expressing the real-time financial data, thereby obtaining the benchmark features of the real-time financial data; Using big data to obtain several key indicators that lead to or affect the benchmark characteristics, randomly extracting a portion of data samples of the same size from each group of real-time financial data, defining the data as a sub-sample data set, and obtaining multiple sub-sample data sets; An agglomerative hierarchical clustering algorithm is introduced, each data in the sub-sample data set is defined as a cluster point, and the Mahalanobis distance matrix between adjacent cluster points is obtained. Based on the Mahalanobis distance matrix, each of the multiple sub-sample data sets is merged and clustered in the agglomerative hierarchical clustering algorithm to obtain a series of new cluster points. According to a series of new cluster points, the corresponding cluster points are returned for iterative update, and the above cluster point merging and clustering and iterative update steps are repeated continuously until all data points are clustered, and an agglomerative hierarchical clustering tree is generated; The leaf node values ​​output by the final hierarchical clustering are obtained through the agglomerative hierarchical clustering tree, and one or more key indicators that affect the current consulting object's execution of financial management purposes are determined based on the leaf node values, and are defined as threatening key indicators.

3. A financial management consulting method based on big data according to claim 2, characterized in that: The step S102 of acquiring historical financial data and real-time financial data related to the financial management purpose based on big data retrieval, and constructing an autonomous encoder and an autonomous decoder to train a sliding window to count the change statistics of the historical financial data to reconstruct the features expressing the real-time financial data, and obtain the benchmark features of the real-time financial data, specifically includes the following steps: Preset the search frequency and established statistical indicators of financial data in big data, and retrieve multiple groups of historical financial data related to financial management purposes within a preset time period in big data according to the search frequency; Setting a sliding window specification based on the search frequency, constructing a sliding window in the plurality of groups of the real-time financial data according to the sliding window specification, and counting the change amount of a predetermined statistical indicator of the data sliding window at an initial position of the historical financial data; After the statistics are completed, the data sliding window is moved to enter the next position in the historical financial data to count the change amount of the predetermined statistical indicator, and the moving data sliding window is repeatedly moved to count the change amount until all data points of the historical financial data are traversed, and finally the historical change statistical value of the predetermined statistical indicator is output; Construct an autonomous encoder model and an autonomous decoder model of a neural network layer architecture, train the historical change statistics of the established statistical indicators through the autonomous encoder model, and generate a mean square error function; Introducing a back propagation algorithm to calculate the mean square error function, obtaining a gradient value of the mean square error function compared to the bias and weight in the autonomous encoder model, iteratively updating the neural network parameters of the autonomous encoder model based on the gradient value, and obtaining a trained autonomous encoder model; Based on big data, multiple groups of real-time financial data for related financial management purposes are obtained, and the trained autonomous encoder model is used to operate on the multiple groups of real-time financial data. The autonomous decoder model is used to decode the low-dimensional financial data features in the autonomous encoder model and reconstruct the expression to obtain the baseline features of the real-time financial data.

4. The financial management consulting method based on big data according to claim 1, characterized in that: The step S104 specifically includes the following steps: Preset a predetermined time interval, obtain N historical financial derivative data generated when the threatening key indicator is within the predetermined time interval, perform mutation analysis on the N historical financial derivative data through the benchmark characteristics of the real-time financial data, and obtain the mutation frequency of the historical financial derivative data; Constructing a Poisson distribution model, estimating a strength threshold of data mutation according to the mutation frequency, performing random variable distribution calculation in the Poisson distribution model based on the strength threshold, and obtaining a probability mass function of mutation of historical financial derivative data; Analyze the global waiting time of sudden changes in historical financial derivative data according to the probability mass function, and construct the global waiting time by using the exponential distribution method to obtain the time interval distribution law of sudden changes in historical financial derivative data; Based on the probability mass function and the time interval distribution law, the evolution trend of the sudden change of historical financial derivative data in the Poisson distribution process is cumulatively described, and the variance value of the probability mass function in the time interval distribution law is obtained during the cumulative evolution trend; Check whether the variance value is greater than or equal to the preset variance value. If so, continue to accumulate the description; if less than, recalculate and accumulate the probability mass function and the time interval distribution law until the variance value is greater than or equal to the preset variance value, and generate a financial mutation trend prediction model; N real-time financial derivative data generated by threatening key indicators within a predetermined time interval are obtained, and the real-time financial derivative data are predicted by a financial mutation trend prediction model to obtain the future financial mutation trend of the threatening key indicators.

5. The financial management consulting method based on big data according to claim 1, characterized in that: The step S106 specifically includes the following steps: Based on big data, a number of historical financial loss data for financial management purposes are obtained within a preset time period under the condition of future financial sudden changes in the threatening key indicators; Calculate the correlation coefficient between the financial management purpose and the future financial sudden change trend of the threatening key indicators based on the historical financial loss data, and establish a correlation coefficient matrix; Constructing a regression equation based on the correlation coefficient matrix, substituting the historical financial loss data into the regression equation to describe the impact relationship between future financial sudden changes and financial management purposes, and generating an impact relationship regression equation; Solving the impact relationship regression equation to obtain a plurality of partial regression coefficients, and determining the impact of the future financial mutation trend of the threatening key indicator on the corresponding financial management purpose according to the partial regression coefficients; Obtaining the intended consulting information on financial management submitted by the current consulting object, extracting the critical loss coefficient that the current consulting object can tolerate through the intended consulting information, and evaluating the critical loss coefficient based on the financial management risk standard to set the risk loss amount range of the current consulting object; Obtain the amount of funds raised by the current consulting object for financial management purposes, and estimate the estimated amount of loss that would be incurred after the amount of funds raised is used for financial management purposes based on the impact range; If the estimated loss amount is not within the risk loss amount range, the estimated loss amount is calibrated as a normal loss; if it is within the risk loss amount range, the estimated loss amount is calibrated as an abnormal loss, and a risk warning signal and risk warning level are generated based on the abnormal loss and uploaded to the financial management consulting terminal to obtain financial management consulting warning information.

6. The financial management consulting method based on big data according to claim 1, characterized in that: The step S108 specifically includes the following steps: Obtaining the lower limit loss amount of the risk loss amount range, calculating the difference between the estimated loss amount of the abnormal loss of the financial management consulting early warning information and the lower limit loss amount, and obtaining the amount deviation value; Obtaining the ideal financial management model of the current consulting object for applying the raised funds amount to financial management purposes through the intended consulting information; wherein the financial management model includes asset allocation, investment portfolio and cost control; Based on the amount deviation value, the application of the loss gradient of the raised funds amount when generating abnormal losses for financial management purposes is gradually optimized until the amount deviation value is eliminated, thereby obtaining the loss optimization parameter; Based on the loss optimization parameters, the allocation of the amount of funds raised in the ideal financial management model is calculated and planned to generate a financial management allocation network, and the financial management allocation network is deployed on the actual financial management model to optimize the use of the amount of funds raised and generate a financial management consulting optimization plan.

7. A financial management consulting system based on big data, characterized in that: The financial management consulting system includes a memory and a processor. The memory stores a financial management consulting method program based on big data. When the financial management consulting method program is executed by the processor, the financial management consulting method steps as described in any one of claims 1 to 6 are implemented.

Images