Business big data analysis method and device based on machine learning and storage medium

Abstract

The invention belongs to the technical field of big data analysis, and provides a machine learning-based business big data analysis method and device and a storage medium, and the method comprises the steps: firstly obtaining the internal data and related external data of an enterprise; calculating a current growth index vector of the enterprise according to the internal data; predicting a growth index change curve of the enterprise according to the current growth index vector, related external data and a pre-constructed hybrid model; according to the method, the current enterprise development is evaluated through the internal data, the subsequent development of the enterprise is predicted in combination with the time sequence model, the influence of external factors on the enterprise development is considered, and the influence is quantized and superposed to the development prediction of the enterprise by adopting the neural network model, so that the predicted growth index change curve is more accurate, and the development prediction efficiency is improved. And meanwhile, the stability and risk of the enterprise are further evaluated, key features of enterprise data are extracted in a deeper level, and a digital strategic reference is provided for enterprise development.

Description

Technical Field

[0001] The present invention belongs to the technical field of big data analysis, and in particular relates to a business big data analysis method, device and storage medium based on machine learning. Background Art

[0002] In the current global economic environment, enterprise digital transformation has become a key factor in promoting economic growth. With the development of technologies such as big data, cloud computing, and artificial intelligence, enterprises are facing the challenge of transforming from traditional operating models to digital and intelligent ones. The calculation of the enterprise digital growth index aims to evaluate the progress and effectiveness of enterprises in the process of digital transformation and provide a reference for enterprises to formulate digital strategies. Through this index, enterprises can better understand their position in digital transformation and formulate corresponding strategies and measures to achieve sustained growth and development.

[0003] In the process of digital transformation, enterprises need to comprehensively consider multiple factors such as strategy, talent, technology, organization, and capital. The existing growth index calculation takes into account relatively one-sided factors, so there is a problem of inaccurate growth index calculation. Summary of the invention

[0004] In view of this, the present invention provides a business big data analysis method, device and storage medium based on machine learning, aiming to solve the problem of inaccurate growth index calculation in the prior art.

[0005] A first aspect of an embodiment of the present invention provides a business big data analysis method based on machine learning, comprising: Obtain the company's internal data and relevant external data; Calculate the company's current growth index vector based on internal data; According to the current growth index vector, relevant external data and pre-built hybrid model, multiple growth index change curves of the enterprise are predicted; wherein the hybrid model includes a time series model and a neural network model in a cascade setting; each growth index change curve corresponds to different relevant external data; Determine the stability and risk of an enterprise based on the growth index change curves of multiple enterprises.

[0006] In a possible implementation, the current growth index vector of the enterprise is calculated based on internal data, including: The internal data is evaluated according to the subjective weighting method and the objective weighting method to obtain the current growth index vector.

[0007] In a possible implementation, the internal data includes financial data, sales data, product data, customer information, and human resources data. The internal data is evaluated according to the subjective weighting method and the objective weighting method to obtain the current growth index vector, including: Obtaining evaluation indicator sets corresponding to various internal data; each evaluation indicator set includes multiple evaluation indicators; According to the pre-built random forest, the correlation between the internal data corresponding to each evaluation indicator; The internal data corresponding to each evaluation index is used as the attribute value of the node, and the association relationship between the internal data corresponding to each evaluation index is used as the attribute value of the edge to establish an evaluation index network; The evaluation index network is evaluated according to the subjective weighting method and the objective weighting method to obtain the current growth index vector.

[0008] In a possible implementation, the evaluation index network is evaluated according to the subjective weighting method and the objective weighting method to obtain the current growth index vector, including: Evaluate the attribute values ​​of the nodes in the evaluation index network according to the subjective weighting method, determine the subjective evaluation values ​​corresponding to each internal data, and form a first vector; Evaluate the attribute values ​​of the nodes in the evaluation index network according to the objective weighting method, determine the objective evaluation values ​​corresponding to each internal data, and form the second vector; Determine the importance parameter based on the attribute value of the edge in the evaluation index network; A current growth index vector is determined according to the first vector, the second vector and the importance parameter.

[0009] In a possible implementation, the hybrid model includes a time series model and a neural network model; based on the current growth index vector, relevant external data and the pre-built hybrid model, multiple growth index change curves of the enterprise are predicted, including: According to the growth index vector and the time series model, the time series of each element in the growth index vector is calculated; Determine multiple parameter groups according to the time series of each element, relevant external data and the neural network model, wherein each parameter group includes multiple external influencing parameters, and each external influencing parameter corresponds to an element in the current growth index vector; According to each parameter group and time series, the growth index change curves of the enterprise are fitted.

[0010] In a possible implementation, the external data includes external data of the industry, external data of upstream industries, and external data of downstream industries; according to the current growth index vector, relevant external data, and the neural network model, multiple parameter groups are determined, including: Determine multiple sets of industry chain influencing parameters based on external data of upstream and downstream industries; Multiple parameter groups are determined based on the time series of each element, external data of the industry, multiple groups of industrial chain influencing parameters and neural network models.

[0011] In a possible implementation, multiple groups of industry chain influencing parameters are determined based on external data of upstream industries and external data of downstream industries, including: Group the upstream industry data and downstream industry data of the enterprise to obtain multiple groups of industry data; Extract features from each group of industry data to obtain the quantitative value of the influencing factors of each group of industry data; The quantitative values ​​of influencing factors are weighted and summed to obtain multiple groups of industrial chain influencing parameters.

[0012] A second aspect of an embodiment of the present invention provides a business big data analysis device based on machine learning, comprising: The acquisition module is used to obtain the internal data and relevant external data of the enterprise; A calculation module is used to calculate the current growth index vector of the enterprise based on internal data; The prediction module is used to predict the growth index change curve of the enterprise based on the current growth index vector, relevant external data and a pre-built hybrid model; wherein the hybrid model includes a time series model and a neural network model in a cascade setting; each growth index change curve corresponds to different relevant external data; Determine the stability and risk of an enterprise based on the growth index change curves of multiple enterprises.

[0013] A third aspect of an embodiment of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the business big data analysis method based on machine learning as described in the first aspect above are implemented.

[0014] A fourth aspect of an embodiment of the present invention provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the business big data analysis method based on machine learning as described in the first aspect above are implemented.

[0015] The business big data analysis method, device and storage medium based on machine learning provided by the embodiment of the present invention first obtain the internal data and relevant external data of the enterprise; then calculate the current growth index vector of the enterprise based on the internal data; then predict the growth index change curve of the enterprise based on the current growth index vector, relevant external data and a pre-built hybrid model; wherein the hybrid model includes a cascaded time series model and a neural network model. The present invention evaluates the current enterprise development through internal data, combines the time series model to predict the subsequent development of the enterprise, and considers the impact of external factors on the development of the enterprise. The neural network model is used to quantify the impact and superimpose it on the development forecast of the enterprise, thereby making the predicted growth index change curve more accurate, and further evaluating the stability and risk of the enterprise, extracting the key features of the enterprise data at a deeper level, and providing a digital strategic reference for enterprise development. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, 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, other drawings can be obtained based on these drawings without paying creative labor.

[0017] Figure 1 It is a flowchart of an implementation of a business big data analysis method based on machine learning provided by an embodiment of the present invention; Figure 2 It is a structural schematic diagram of a business big data analysis method and device based on machine learning provided by an embodiment of the present invention; Figure 3 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0019] Figure 1 : is a flowchart of the implementation of the business big data analysis method based on machine learning provided by an embodiment of the present invention. Figure 1 As shown, in some embodiments, a business big data analysis method based on machine learning includes: S110, obtaining the internal data and relevant external data of the enterprise; S120, calculating the enterprise's current growth index vector based on internal data; S130, predicting the growth index change curve of the enterprise based on the current growth index vector, relevant external data and a pre-built hybrid model; wherein the hybrid model includes a cascaded time series model and a neural network model, and each growth index change curve corresponds to different relevant external data.

[0020] S140, determines the stability and risk of an enterprise based on the growth index change curves of multiple enterprises.

[0021] In the embodiment of the present invention, the internal data of the enterprise includes financial data, sales data, product data, customer information, and human resources data; Among them, financial data is information that reflects the financial status and operating results of an enterprise. It covers the company's assets, liabilities, owner's equity, income, costs, expenses, profits and other aspects. By analyzing financial data, corporate managers can understand the company's debt repayment ability (such as current ratio, quick ratio and other indicators), profitability (such as gross profit margin, net profit margin, etc.) and operating ability (such as inventory turnover, accounts receivable turnover, etc.). Investors and creditors also rely on financial data to assess the value and risk of a company. For example, investors will decide whether to invest based on the company's profitability and asset-liability structure; creditors will determine whether to provide loans and the amount and interest rate of the loan based on debt repayment ability indicators.

[0022] Sales data reflects the relevant information of the company's sales activities. It includes sales volume, sales channel data, sales area data, customer purchase frequency, order amount, etc. By analyzing the trend of sales volume and sales volume, we can determine the life cycle stage of the product, such as growth, maturity or decline. It is instructive for the formulation and adjustment of sales strategies. For example, if the sales share of a certain sales channel continues to decline, the company may need to optimize the channel or increase investment in other channels.

[0023] Product data involves various attribute information of products, including product specifications (such as size, weight, performance parameters, etc.), product quality data (such as qualified rate, defective rate, etc.), product R&D data (such as R&D investment, R&D cycle, etc.), product inventory data, etc. By analyzing product quality data and customer feedback, we can find problems with the product and improve product design and production processes. This affects product inventory management and supply chain decisions. Product inventory data can help companies arrange production plans reasonably and avoid inventory backlogs or out-of-stock phenomena.

[0024] Customer information includes basic customer information (such as name, contact information, address, etc.), customer purchase history (such as purchased products or services, purchase amount, purchase time, etc.), customer preferences (such as favorite product color, function, etc.), customer satisfaction survey data, etc. By analyzing customer purchase history and preferences, companies can segment customers and provide personalized product recommendations and marketing activities for different customer groups. This helps improve customer satisfaction and loyalty. Based on customer satisfaction survey data, companies can identify deficiencies in their services and improve them, thereby enhancing customer trust and reliance on the company.

[0025] Human resource data reflects relevant information about the company's employees. It includes basic employee information (such as age, gender, education, etc.), employee attendance data, employee performance data, employee salary data, employee training data, etc. By analyzing employee performance data, decisions on employee promotion, reward or dismissal can be made; salary data helps companies formulate a reasonable salary system. It helps companies plan human resource development strategies. For example, based on the age structure and education distribution of employees, companies can determine the direction of future talent recruitment and training.

[0026] In an embodiment of the present invention, relevant external data of an enterprise can be obtained through the following methods: government public data, industry reports, social media data, financial market data, etc.

[0027] The time series model makes predictions based on the time trend and seasonal characteristics of the historical growth index data. For example, if the ARIMA model is used, it will analyze the autocorrelation, differential stationarity and other characteristics of the growth index. The neural network model takes external data as input and evaluates the impact of external data on each element of the growth index vector through its complex network structure and learning mechanism. Therefore, in an embodiment of the present invention, the prediction results output by the time series model are input into the neural network model, and are calculated again in combination with the external data to form a hybrid model with a cascade setting, thereby improving the accuracy of the growth index measurement.

[0028] In some embodiments, calculating the current growth index vector of the enterprise based on internal data includes: evaluating the internal data based on a subjective weighting method and an objective weighting method to obtain the current growth index vector.

[0029] In the embodiment of the present invention, the subjective weighting method may be a hierarchical analysis method, a principal component analysis method, etc., which is not limited here. The objective weighting method may be a coefficient of variation method, an entropy weight method, etc., which is not limited here. The evaluation results of the subjective weighting method and the objective weighting method are weighted and summed to obtain the current growth index vector.

[0030] In some embodiments, the internal data includes financial data, sales data, product data, customer information, and human resources data; the internal data is evaluated according to subjective weighting methods and objective weighting methods to obtain a current growth index vector, including: obtaining an evaluation indicator set corresponding to each internal data; each evaluation indicator set includes multiple evaluation indicators; according to a pre-constructed random forest, the correlation between the internal data corresponding to each evaluation indicator; using the internal data corresponding to each evaluation indicator as the attribute value of the node, and the correlation between the internal data corresponding to each evaluation indicator as the attribute value of the edge, to establish an evaluation indicator network; evaluating the evaluation indicator network according to the subjective weighting method and the objective weighting method to obtain the current growth index vector.

[0031] In the embodiment of the present invention, since the internal data is relatively complex, and the financial data, sales data, product data, customer information, and human resources data of different types of companies are all different, and the emphasis and attention on these types of data are also different, the present invention sets an evaluation index set for each internal data, and each internal data needs to be sorted and calculated according to each evaluation index in the index set. For example, the original data such as financial statements in financial data need to be sorted and calculated into profitability indicators, debt repayment ability indicators, operating ability indicators, etc.

[0032] In an embodiment of the present invention, after calculating the values ​​of these evaluation indicators based on internal data, considering that these data do not have an isolated impact on the development of the enterprise, but are a process of mutual influence, the present invention also uses random forests to evaluate the correlation between the various evaluation indicators, and then uses the indicator values ​​as the attribute values ​​of the network nodes, connects these network nodes to each other, forms a three-dimensional indicator evaluation network, and adds the correlation relationship as the attribute value of the edge to the network.

[0033] In some embodiments, the evaluation index network is evaluated according to a subjective weighting method and an objective weighting method to obtain a current growth index vector, including: evaluating the attribute values ​​of the nodes in the evaluation index network according to the subjective weighting method, determining the subjective evaluation values ​​corresponding to each internal data, and forming a first vector; evaluating the attribute values ​​of the nodes in the evaluation index network according to the objective weighting method, determining the objective evaluation values ​​corresponding to each internal data, and forming a second vector; determining importance parameters according to the attribute values ​​of the edges in the evaluation index network; determining the current growth index vector according to the first vector, the second vector and the importance parameters.

[0034] In the embodiment of the present invention, each internal data is evaluated subjectively and objectively, and the evaluation values ​​are organized into vectors, namely the first vector and the second vector mentioned above, the two vectors have the same scale, and the elements correspond one to one. Usually, the corresponding values ​​of the two vectors are weighted and summed according to the preset weights to obtain the current growth index vector.

[0035] In addition, the present invention takes into account the association between various data, and adds the association relationship as the attribute value of the edge to the evaluation network. The sum of the attribute values ​​of the edges corresponding to each internal data, divided by the sum of the attribute values ​​of all edges in the evaluation index network, is the importance parameter of the financial data. For a certain internal data, such as financial data, the sum of the attribute values ​​of all the edges of the financial data is calculated, and then divided by the sum of the attribute values ​​of all the edges in the evaluation index network, which is the importance parameter of the financial data. That is, the stronger the association between a certain data and other data, the greater the importance parameter.

[0036] In an embodiment of the present invention, after calculating the importance parameter, the first vector and the second vector need to adjust each element in the vector according to the importance parameter of each element. Specifically, a preset interval of importance can be set. If the importance parameter is greater than the upper limit of the interval, the value of the element is increased. If it is within the preset interval, it remains unchanged. If it is less than the lower limit of the preset interval, the value of the element is reduced. For example, financial data corresponds to an element a1 and b1 in the first vector and the second vector respectively, and its importance is greater than the upper limit of the preset interval. The corresponding elements in the first vector and the second vector are increased to a1+a0 and b1+b0. Among them, a0 and b0 are both preset values. The adjusted first vector and the second vector are weighted and summed to obtain the current growth index vector.

[0037] In some embodiments, the hybrid model includes a time series model and a neural network model; based on the current growth index vector, relevant external data and a pre-built hybrid model, the growth index change curve of the enterprise is predicted, including: based on the growth index vector and the time series model, the time series of each element in the growth index vector is calculated; based on the time series of each element, relevant external data and the neural network model, multiple parameter groups are determined, wherein each parameter group includes multiple external influencing parameters, and each external influencing parameter corresponds to an element in the current growth index vector; based on multiple parameter groups and time series, each growth index change curve of the enterprise is fitted.

[0038] In an embodiment of the present invention, after the growth index vector [a1, ..., an] is input into the time series model, each element is predicted respectively to obtain multiple time series [a1t0, a1t1, ..., a1tm], ..., [ant0, ant1, ..., antm]. Each time series needs to be merged into a vector to reflect the development of the enterprise. For example, a1t0, ..., ant0 are weighted summed, and so on, until a1tm, ..., antm are weighted summed to obtain a merged vector, and then the discrete vector is curve fitted to obtain the growth index change curve of the enterprise.

[0039] In this process, relevant external data determines the weights in the fitting process. Under different industries, different policies, and different industry environments, the needs of enterprise development are different. For example, when the industry is developing well and the policy is positively oriented, appropriate increase in investment in R&D and personnel training is conducive to enterprise development. Expenditure should be appropriately increased, and short-term deterioration of financial data should be allowed. The corresponding element values ​​of these two aspects in the growth index vector are more important, that is, increase the elements of these two items and reduce the elements corresponding to financial data.

[0040] Therefore, different relevant external data will be used, the weights obtained will be different, and the fitted curves will also be different. Different growth index change curves can reflect the development status of the company under different external influences.

[0041] In some embodiments, the external data includes external data of the industry, external data of upstream industries and external data of downstream industries; multiple parameter groups are determined based on the time series of each element, relevant external data and neural network model, including: determining the influencing parameters of the industrial chain based on the external data of upstream industries and external data of downstream industries; multiple parameter groups are determined based on the time series of each element, external data of the industry, influencing parameters of the industrial chain and neural network model.

[0042] In addition to paying attention to the industry itself, the development of an enterprise also needs to pay attention to the status of upstream and downstream industries to a certain extent in order to cope with possible external changes. Therefore, when calculating external influencing parameters, the present invention calculates the influencing parameters of the industrial chain through the external data of the upstream and downstream industries, and inputs them into the neural network model to jointly calculate the external influencing parameters.

[0043] In some embodiments, multiple groups of industrial chain influencing parameters are determined based on upstream industry external data and downstream industry external data, including: grouping the upstream industry data and downstream industry data of the enterprise to obtain multiple groups of industry data; extracting features from each group of industry data to obtain quantitative values ​​of influencing factors for each group of industry data; and performing weighted summation on the quantitative values ​​of influencing factors to obtain industrial chain influencing parameters.

[0044] In the embodiment of the present invention, the features that can be extracted from the external data of the upstream industry include: price features, supply stability features, technological innovation features, and policy and regulation features. The features that can be extracted from the external data of the downstream industry include: market demand features, market competition features, etc.

[0045] By normalizing the extracted features, we can get the quantitative value of the influencing factors of each feature, and then perform weighted summation according to the pre-set weights to get the industry chain impact parameters.

[0046] Each enterprise has multiple upstream industries and multiple downstream industries. For example, lithium battery manufacturing enterprises, their upstream industries include suppliers of various minerals / compounds such as lithium, cobalt, nickel, and manganese, as well as manufacturers of various materials such as positive electrodes, negative electrodes, diaphragms, and electrolytes, as well as many research and development and testing industries. Its downstream industries include manufacturers of new energy vehicles, consumer electronics, energy storage systems, and smart wearable devices. It can be seen that the industrial chain of an enterprise is complex, and it is difficult to describe the specific impact of the industrial chain on the enterprise in a unified analysis method. Therefore, the present invention clusters the data of each upstream industry and downstream industry, groups them according to the clustering results, extracts features from each group of data, and then quantifies and weights them to obtain different industrial chain impact parameters.

[0047] The growth index change curves of enterprises calculated based on different industrial chain influencing parameters can reflect the impact of different external changes on the development of the enterprise.

[0048] In some embodiments, the stability and risk of an enterprise are determined based on multiple growth index change curves of the enterprise, including: fitting a first curve based on each growth index change curve; calculating the mean and variance of the distance between each growth index change curve and the first curve, and simultaneously calculating the trend reversal index of each growth index change curve; determining the stability and risk of the enterprise based on the trend reversal index, mean and variance.

[0049] Compare the stability of the growth index change curve under different external influences (reflected by different industry chain influencing parameters). Observe the fluctuation of the enterprise growth index change curve and the degree of change in the long-term trend when external data (such as upstream raw material price fluctuations, downstream market demand changes, etc.) changes.

[0050] If the fluctuation range and trend changes of the growth index curve of an enterprise are small under various external changes, it means that the enterprise has strong stability and can effectively cope with interference from the external environment.

[0051] For example, a technology company, facing the dual external pressures of tight upstream chip supply and the launch of new products by downstream competitors, the fluctuations of its growth index change curve are still within a controllable range, and the long-term trend has not deteriorated significantly, which reflects the company's stability in a complex external environment.

[0052] The mean can reflect the long-term trend of the growth index change curve to a certain extent, while the variance can reflect the fluctuation of the growth index change curve to a certain extent. Therefore, by taking the weighted sum of the calculated mean and variance, we can get the stability score of the enterprise in a complex external environment.

[0053] In addition, it is also necessary to check whether there is a trend reversal in each growth index change curve, that is, a sudden change from an upward trend to a downward trend or vice versa. The larger the absolute value of the slope after the reversal, the higher the risk of trend reversal. At the same time, analyze the external factors that lead to trend reversal, such as the surge in upstream raw material prices and the loss of major downstream customers, and determine the contribution of each factor to the trend reversal in combination with the industry chain influencing parameters.

[0054] For example, for a lithium battery manufacturer, when the subsidy policy for new energy vehicles suddenly declined (external factors), the downward trend of its growth index change curve reversed. By analyzing the influencing parameters of the industrial chain, it can be seen that the reduction in demand from downstream automobile manufacturers has had a significant negative impact on the company's growth, thereby determining a higher trend reversal risk.

[0055] Therefore, the average rate of change of the curve after each trend reversal is calculated as the trend reversal index. The weighted sum of the mean and variance of the trend reversal index can be used to obtain the risk score of the enterprise in a complex external environment.

[0056] In addition, internal data and external data are also associated with tourism management data. For example, the internal financial data of tourism enterprises, such as tourism project revenue and expenditure, assets and liabilities, can reflect profitability and debt repayment ability. Sales data covers the number of tourists, tourism product sales and the proportion of each channel, which helps to grasp market demand and channel advantages and disadvantages. Product data involves product characteristics such as travel routes and hotel facilities and customer feedback, which is conducive to product improvement. Customer information includes tourists' age, region, preferences, consumption frequency and satisfaction, which can achieve precision marketing and service improvement. Human resources data covers employees' professional background, service performance, and training investment, which is related to service quality and corporate development strategy. For example, a travel agency analyzes customer information and accurately launches special products such as parent-child tours and elderly health tours, which greatly increases sales and customer satisfaction. Obtaining policies and regulations, scenic spot ratings, and tourism statistics from government tourism departments, such as policy support, is conducive to planning business directions. Industry reports provide market size, growth rate, competitive landscape and trend information, which helps to position and formulate competitive strategies. Social media data mining tourists' evaluations, sharing and hot topics can provide insights into market preferences and reputation. In financial market data, exchange rates and interest rates affect outbound travel costs and investment decisions, and tourism-related stock indices reflect the industry's prosperity. For example, before the peak tourist season, timely optimize popular routes based on social media popularity, effectively improving tourist experience and corporate reputation.

[0057] In addition, tourism management data will also have an impact on the above-mentioned time series model, and thus affect the calculation of the growth index change curve. In the time series model, periodic factors such as tourism off-season and holidays are taken into account, and historical data is used to accurately predict the time trend of growth index elements. The model constructs an input layer based on the characteristics of the tourism market, classifies external data into inputs, and accurately quantifies the weights of the impact on the growth index elements, such as the determination of the weights of the impact of the upgrading of supporting facilities around scenic spots on related tourism enterprises. In order to accurately reflect the tourism business, the weights and correlations of the internal data evaluation indicator set are re-evaluated. For example, for hotel companies, the weights of the room occupancy rate indicators are dynamically adjusted according to the off-season and peak season, and the growth index vector calculation is optimized. In this way, the integration of all-round tourism management data can greatly improve the accuracy and applicability of the enterprise digital development index calculation, and strongly support the strategic decision-making of the enterprise.

[0058] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0059] Figure 2 Schematic diagram of the structure of the business big data analysis method and device based on machine learning provided by an embodiment of the present invention. Figure 2 As shown, in some embodiments, a business big data analysis method device 2 based on machine learning includes: An acquisition module 210 is used to acquire internal data and related external data of an enterprise; A calculation module 220, for calculating the current growth index vector of the enterprise based on internal data; The prediction module 230 is used to predict multiple growth index change curves of the enterprise according to the current growth index vector, relevant external data and a pre-built hybrid model; wherein the hybrid model includes a cascaded time series model and a neural network model; each growth index change curve corresponds to different relevant external data; The evaluation module 240 is used to determine the stability and risk of an enterprise based on a plurality of growth index change curves of the enterprises.

[0060] Optionally, the evaluation module 240 is used to fit the first curve based on each growth index change curve; calculate the mean and variance of the distance between each growth index change curve and the first curve, and simultaneously calculate the trend reversal index of each growth index change curve; determine the stability and risk of the enterprise based on the trend reversal index, mean and variance.

[0061] Optionally, the calculation module 220 is used to evaluate the internal data according to the subjective weighting method and the objective weighting method to obtain the current growth index vector.

[0062] Optionally, the internal data includes financial data, sales data, product data, customer information, and human resources data; the calculation module 220 is used to obtain an evaluation index set corresponding to each internal data; each evaluation index set includes multiple evaluation indicators; according to a pre-constructed random forest, the correlation between the internal data corresponding to each evaluation indicator; the internal data corresponding to each evaluation indicator is used as the attribute value of the node, and the correlation between the internal data corresponding to each evaluation indicator is used as the attribute value of the edge to establish an evaluation index network; the evaluation index network is evaluated according to the subjective weighting method and the objective weighting method to obtain the current growth index vector.

[0063] Optionally, the calculation module 220 is used to evaluate the attribute values ​​of the nodes in the evaluation index network according to the subjective weighting method, determine the subjective evaluation values ​​corresponding to each internal data, and form a first vector; evaluate the attribute values ​​of the nodes in the evaluation index network according to the objective weighting method, determine the objective evaluation values ​​corresponding to each internal data, and form a second vector; determine the importance parameter according to the attribute values ​​of the edges in the evaluation index network; determine the current growth index vector according to the first vector, the second vector and the importance parameter.

[0064] Optionally, the hybrid model includes a time series model and a neural network model; the prediction module 230 is used to calculate the time series of each element in the growth index vector based on the growth index vector and the time series model; determine multiple parameter groups based on the time series of each element, relevant external data and the neural network model, wherein each parameter group includes multiple external influencing parameters, and each external influencing parameter corresponds to an element in the current growth index vector; and fit the growth index change curve of the enterprise based on each parameter group and the time series.

[0065] Optionally, the external data includes external data of the industry, external data of upstream industries and external data of downstream industries; the prediction module 230 is used to determine multiple groups of industrial chain influencing parameters based on the external data of upstream industries and downstream industries; determine the external influencing parameters of each element in the current growth index vector based on the time series of each element, external data of the industry, multiple groups of industrial chain influencing parameters and a neural network model.

[0066] Optionally, the prediction module 230 is used to group the upstream industry data and downstream industry data of the enterprise to obtain multiple groups of industry data; perform feature extraction on each group of industry data to obtain quantitative values ​​of influencing factors of each group of industry data; perform weighted summation on the quantitative values ​​of influencing factors to obtain multiple groups of industry chain influencing parameters.

[0067] The business big data analysis method device based on machine learning provided in this embodiment can be used to execute the above method embodiments. Its implementation principles and technical effects are similar, and this embodiment will not be repeated here.

[0068] Figure 3 Schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. Figure 3 As shown, an electronic device 3 provided by an embodiment of the present invention includes: a processor 30, a memory 31, and a computer program 32 stored in the memory 31 and executable on the processor 30. When the processor 30 executes the computer program 32, the steps in the above-mentioned output voltage calibration method based on LLC are implemented, for example Figure 2 Alternatively, when the processor 30 executes the computer program 32, the functions of each module / unit in the above-mentioned system embodiments are implemented, for example Figure 3 The functions of each module are shown.

[0069] Exemplarily, the computer program 32 may be divided into one or more modules / units, one or more modules / units are stored in the memory 31, and are executed by the processor 30 to implement the present invention. One or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the computer program 32 in the electronic device 3.

[0070] The electronic device 3 may be a mobile phone, MCU, ECU, industrial computer, etc., which is not limited here. The electronic device 3 may include, but is not limited to, a processor 30 and a memory 31. Those skilled in the art will understand that Figure 3 It is only an example of the electronic device 3 and does not constitute a limitation of the electronic device 3. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the electronic device may also include input and output devices, network access devices, buses, etc.

[0071] The processor 30 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0072] The memory 31 may be an internal storage unit of the electronic device 3, such as a hard disk or memory of the electronic device 3. The memory 31 may also be an external storage device of the electronic device 3, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 3. Further, the memory 31 may also include both an internal storage unit and an external storage device of the electronic device 3. The memory 31 is used to store computer programs and other programs and data required by the electronic device. The memory 31 may also be used to temporarily store data that has been output or is to be output.

[0073] An embodiment of the present invention provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps in the above-mentioned LLC-based output voltage calibration method embodiment are implemented.

[0074] The computer readable storage medium stores a computer program 32, which includes program instructions. When the program instructions are executed by the processor 30, all or part of the processes in the above-mentioned embodiment method are implemented. The computer program 32 can also be used to instruct related hardware to complete. The computer program 32 can be stored in a computer readable storage medium. When the computer program 32 is executed by the processor 30, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program 32 includes computer program code, which can be in the form of source code, object code, executable file or some intermediate form. The computer readable medium can include: any entity or device that can carry computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium.

[0075] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A business big data analysis method based on machine learning, characterized in that: include: Acquire business big data information; wherein the business big data information includes internal data and relevant external data of the enterprise; Calculate the current growth index vector of the enterprise based on the internal data; Predicting multiple growth index change curves of the enterprise according to the current growth index vector, the relevant external data and the pre-built hybrid model; wherein the hybrid model includes a cascaded time series model and a neural network model; each growth index change curve corresponds to different relevant external data; The stability and risk of the enterprises are determined based on the growth index change curves of the multiple enterprises.

2. The business big data analysis method based on machine learning according to claim 1 is characterized in that: Determine the stability and risk of the enterprise based on the multiple growth index change curves of the enterprise, including: According to each growth index change curve, a first curve is obtained by fitting; Calculate the mean and variance of the distance between each growth index change curve and the first curve, and simultaneously calculate the trend reversal index of each growth index change curve; The stability and risk of an enterprise are determined based on the trend reversal index, the mean and the variance.

3. The business big data analysis method based on machine learning according to claim 1 is characterized in that: Based on the internal data, the current growth index vector of the enterprise is calculated, including: The internal data is evaluated according to the subjective weighting method and the objective weighting method to obtain the current growth index vector.

4. The business big data analysis method based on machine learning according to claim 3 is characterized in that: The internal data includes financial data, sales data, product data, customer information, and human resources data; The internal data is evaluated according to the subjective weighting method and the objective weighting method to obtain the current growth index vector, including: Obtaining evaluation indicator sets corresponding to various internal data; each evaluation indicator set includes multiple evaluation indicators; According to the pre-built random forest, the correlation between the internal data corresponding to each evaluation indicator; The internal data corresponding to each evaluation index is used as the attribute value of the node, and the association relationship between the internal data corresponding to each evaluation index is used as the attribute value of the edge to establish an evaluation index network; The evaluation index network is evaluated according to a subjective weighting method and an objective weighting method to obtain the current growth index vector.

5. The business big data analysis method based on machine learning according to claim 4 is characterized in that: The evaluation index network is evaluated according to the subjective weighting method and the objective weighting method to obtain the current growth index vector, including: Evaluate the attribute values ​​of the nodes in the evaluation index network according to the subjective weighting method, determine the subjective evaluation values ​​corresponding to each internal data, and form a first vector; Evaluate the attribute values ​​of the nodes in the evaluation index network according to the objective weighting method, determine the objective evaluation values ​​corresponding to each internal data, and form the second vector; Determine the importance parameter based on the attribute value of the edge in the evaluation index network; The current growth index vector is determined according to the first vector, the second vector and the importance parameter.

6. The business big data analysis method based on machine learning according to claim 1 is characterized in that: According to the current growth index vector, the relevant external data and the pre-built hybrid model, multiple growth index change curves of the enterprise are predicted, including: Calculating the time series of each element in the growth index vector according to the growth index vector and the time series model; Determine a plurality of parameter groups according to the time series of each element, the relevant external data and the neural network model, wherein each parameter group includes a plurality of external influencing parameters, and each external influencing parameter corresponds to an element in the current growth index vector; According to each parameter group and the time series, each growth index change curve of the enterprise is obtained by fitting.

7. The business big data analysis method based on machine learning according to claim 6 is characterized in that: The external data includes external data of the industry, external data of upstream industries and external data of downstream industries; according to the current growth index vector, the relevant external data and the neural network model, multiple parameter groups are determined, including: Determine multiple sets of industry chain influencing parameters based on external data of upstream and downstream industries; Multiple parameter groups are determined according to the time series of each element, the external data of the industry, the multiple groups of industrial chain influencing parameters and the neural network model.

8. The business big data analysis method based on machine learning according to claim 7 is characterized in that: Based on external data of upstream and downstream industries, multiple sets of industry chain influencing parameters are determined, including: Group the upstream industry data and downstream industry data of the enterprise to obtain multiple groups of industry data; Extract features from each group of industry data to obtain the quantitative values ​​of the influencing factors of each group of industry data; The quantified values ​​of the influencing factors are weighted and summed to obtain the multiple groups of industrial chain influencing parameters.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the business big data analysis method based on machine learning as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the business big data analysis method based on machine learning as described in any one of claims 1 to 7 above are implemented.

Images