Oil and gas enterprise financial data multi-dimensional risk assessment method, device and equipment

Abstract

The invention discloses an oil and gas enterprise financial data multi-dimensional risk assessment method, device and equipment, and the method comprises the steps: obtaining the current financial data of an oil and gas enterprise, and constructing a high-dimensional index system of underlying report data based on the related data of three reports in the current financial data; wherein the high-dimension index system comprises general financial indexes and industry characteristic financial indexes; and inputting the high-dimensional index system of the underlying report data into a pre-trained multi-dimensional risk assessment model, and determining a multi-dimensional risk assessment value of the financial data of the oil and gas enterprise. According to the method, the financial analysis efficiency and accuracy are improved, accurate benchmarking and competitive advantage evaluation are realized, resource allocation and strategic decision are optimized, the financial risk management capability is further enhanced, and financial management innovation of the oil and gas industry is promoted.

Description

A method, apparatus, and equipment for multi-dimensional risk assessment of financial data of oil and gas enterprises. Technical Field

[0001] This invention relates to the fields of artificial intelligence and computer technology, and in particular to a method, apparatus and equipment for multi-dimensional risk assessment of financial data of oil and gas enterprises. Background Technology

[0002] In certain industries, such as the oil and gas sector, the financial condition of a group's subsidiaries directly reflects the company's operational efficiency, resource allocation, and development potential. Accurate and comprehensive analysis of subsidiary financial data plays an indispensable role in the group's strategic decision-making, resource allocation, and overall competitiveness enhancement.

[0003] Traditional financial data analysis methods often rely on manual processing and simple statistical tools, which have the following obvious shortcomings:

[0004] Low data processing efficiency: Faced with large amounts of historical financial data, manual processing is slow, prone to errors, and difficult to complete multi-dimensional analysis in a short period of time. For example, an oil and gas subsidiary generates a large amount of financial statement data every month. If it relies on manual verification and calculation, it may take several weeks or even months.

[0005] Limited analytical dimensions: Traditional methods typically only perform simple financial indicator calculations, such as revenue and profit, failing to delve into the underlying relationships and trends within the data. For example, in cost analysis, traditional methods may focus solely on direct costs, neglecting the impact of indirect and implicit costs on a company's financial situation.

[0006] Lack of effective comparison with peers: It is difficult to obtain financial data from companies of similar size in the same industry for benchmarking analysis, making it impossible to accurately assess one's own position and competitive advantages within the industry. For example, a newly established oil and gas subsidiary wants to understand the gap between its R&D investment and that of leading companies in the industry, but due to a lack of effective data channels and analytical methods, it is unable to make an accurate comparison. Summary of the Invention

[0007] To improve the efficiency and accuracy of financial analysis, achieve precise benchmarking and competitive advantage assessment, optimize resource allocation and strategic decision-making, further enhance financial risk management capabilities, and promote financial management innovation in the oil and gas industry, this invention provides a method, apparatus, and equipment for multi-dimensional risk assessment of financial data for oil and gas enterprises.

[0008] In a first aspect, embodiments of the present invention provide a multi-dimensional risk assessment method for financial data of oil and gas enterprises, which may include:

[0009] Obtain the current financial data of oil and gas companies, and based on the relevant data of the three major financial statements in the current financial data, construct a high-dimensional indicator system for the underlying financial data; wherein, the high-dimensional indicator system includes general financial indicators and industry-specific financial indicators;

[0010] By inputting the high-dimensional indicator system of the underlying report data into a pre-trained multi-dimensional risk assessment model, the risk assessment values ​​of the financial data of oil and gas companies are determined.

[0011] In one embodiment, the multidimensional risk assessment model is pre-trained using the following method:

[0012] Obtain a training sample set, wherein each sample in the training sample set includes a different dimension indicator system and its risk preference value; wherein the risk preference value is set based on the experience of financial experts in the oil and gas industry;

[0013] The LightGBM model is trained using samples from the training sample set. The different dimensional indicator systems in the samples are passed through the correlation layer of the LightGBM model to obtain the risk preference values ​​corresponding to the different dimensional indicator systems in the samples. The parameters of the LightGBM model are then estimated to obtain a multi-dimensional risk assessment model.

[0014] In another embodiment, obtaining the training sample set may include:

[0015] Obtain relevant data from the three major financial statements of oil and gas companies in their historical financial data.

[0016] Based on the selection of general financial indicators and industry-specific financial indicators for various dimensions of oil and gas enterprise finance by industry experts;

[0017] Based on the relevant data from the three major financial statements in historical financial data, and using general financial indicators and industry-specific financial indicators of various dimensions as the data architecture, a high-dimensional indicator system based on the underlying financial statement data is formed for the historical financial data of oil and gas companies.

[0018] In another embodiment, the general financial metrics include: net assets, debt-to-equity ratio, net profit margin, return on total assets, return on equity, accounts receivable turnover, inventory turnover, revenue growth rate, and / or total interest growth rate.

[0019] The industry-specific financial indicators include: technology investment ratio, guarantees exceeding equity ratio, guarantees exceeding the group's guarantee plan, guarantees outside the group, budget overrun indicators, and / or centralized settlement rate.

[0020] In another embodiment, after obtaining the training sample set, data cleaning may also be performed on each sample in the training sample set.

[0021] Secondly, embodiments of the present invention provide a method for training a multi-dimensional risk assessment model, which may include:

[0022] Obtain a training sample set, wherein each sample in the training sample set includes a different dimension indicator system and its risk preference value; wherein the risk preference value is set based on the experience of financial experts in the oil and gas industry;

[0023] The LightGBM model is trained using samples from the training sample set. The different dimensional indicator systems in the samples are passed through the correlation layer of the LightGBM model to obtain the risk preference values ​​corresponding to the different dimensional indicator systems in the samples. The parameters of the LightGBM model are then estimated to obtain a multi-dimensional risk assessment model.

[0024] Thirdly, embodiments of the present invention provide a multi-dimensional risk assessment device for financial data of oil and gas enterprises, which may include:

[0025] The first acquisition module is used to acquire the current financial data of oil and gas companies;

[0026] The construction module is used to build a high-dimensional indicator system for the underlying financial data based on the relevant data of the three major financial statements in the current financial data; wherein, the high-dimensional indicator system includes general financial indicators and industry-specific financial indicators;

[0027] The determination module is used to input the high-dimensional indicator system of the underlying report data into the pre-trained multi-dimensional risk assessment model to determine the multi-dimensional risk assessment values ​​of the financial data of oil and gas companies.

[0028] Fourthly, embodiments of the present invention provide a multi-dimensional risk assessment model training device, which may include:

[0029] The second acquisition module is used to acquire a training sample set, wherein each sample in the training sample set includes a different dimension indicator system and its risk preference value; wherein the risk preference value is set based on the experience of financial experts in the oil and gas industry;

[0030] The training module is used to train the LightGBM model with samples from the training sample set. The different dimension index systems in the samples are passed through the correlation layer of the LightGBM model to obtain the risk preference values ​​corresponding to the different dimension index systems in the samples, so as to perform parameter estimation on the LightGBM model to obtain a multi-dimensional risk assessment model.

[0031] Fifthly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon. When executed by a processor, the program implements the multi-dimensional risk assessment method for financial data of oil and gas enterprises as described in the first aspect, or implements the multi-dimensional risk assessment model training method as described in the second aspect.

[0032] In a sixth aspect, embodiments of the present invention provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the multi-dimensional risk assessment method for financial data of oil and gas enterprises as described in the first aspect, or implements the multi-dimensional risk assessment model training method as described in the second aspect.

[0033] The beneficial effects of the above-described technical solutions provided in the embodiments of the present invention include at least the following:

[0034] This invention provides a method, apparatus, and equipment for multi-dimensional risk assessment of financial data for oil and gas enterprises. Firstly, it improves the quality of financial analysis, reducing errors and omissions caused by manual analysis and ensuring the consistency and reliability of data interpretation. Secondly, it enhances the scientific nature of decision-making, providing strong support for the group's strategic decisions based on accurate financial analysis and benchmarking results, improving the scientific and rational nature of decisions; it helps the group predict market trends and potential risks in advance, and adjust strategies in a timely manner to cope with changes. Thirdly, it promotes competition and development among subsidiaries, allowing subsidiaries to understand their own gaps and advantages compared to peers, clarifying improvement directions, and stimulating their motivation to improve financial performance and competitiveness; it promotes healthy competition among subsidiaries, jointly improving the overall financial performance of the group. Fourthly, it enhances the industry's position, enabling the oil and gas group to be at the forefront of financial management in the industry, enhancing its influence and competitiveness; it provides a reference for other companies in the same industry, promoting innovation and development in financial management throughout the oil and gas industry. Fifthly, it strengthens risk management, promptly identifying potential financial risks, such as liquidity risks and debt risks, and taking corresponding preventive measures; it ensures the stability and sustainable development of the group's finances.

[0035] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0036] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0037] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0038] Figure 1 is a flowchart of the multi-dimensional risk assessment model training method provided in an embodiment of the present invention;

[0039] Figure 2 is a schematic diagram of the structure of the multi-dimensional risk assessment model training device provided in an embodiment of the present invention;

[0040] Figure 3 is a flowchart of the multi-dimensional risk assessment method for financial data of oil and gas enterprises provided in the embodiment of the present invention;

[0041] Figure 4 is a schematic diagram of the structure of the multi-dimensional risk assessment device for financial data of oil and gas enterprises provided in an embodiment of the present invention. Detailed Implementation

[0042] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0043] After conducting research and analysis, the inventors identified shortcomings in existing technologies regarding modeling and large-scale model applications. Specifically, in the field of financial data modeling, existing modeling methods and technologies face numerous problems when processing complex financial data from the oil and gas industry.

[0044] Poor model adaptability: General modeling methods are difficult to adapt to the unique financial structure and business model of the oil and gas industry. For example, R&D investment in the oil and gas industry is characterized by long cycles and high risks, and existing financial models may not be able to accurately assess its impact on the company's future financial situation.

[0045] Data integration is challenging: Inconsistent data formats and standards among oil and gas subsidiaries make data integration difficult, impacting the accuracy and reliability of modeling. For example, different subsidiaries use different financial software, resulting in varying data fields and coding rules, making data integration exceptionally complex.

[0046] Large-scale model applications are still in their infancy: While large-scale models have achieved some success in other fields, their application in financial analysis within the oil and gas industry is still in its early stages, lacking mature case studies and experience. For example, when applying large-scale models to financial forecasting in the oil and gas industry, the training and optimization of the models face numerous challenges due to the high sensitivity and confidentiality requirements of the data.

[0047] In summary, existing technologies have many shortcomings in the analysis of financial data of subsidiaries in the oil and gas industry, failing to meet the group's needs for refined management and strategic decision-making. Therefore, there is an urgent need for an innovative solution to achieve efficient modeling and in-depth analysis of historical financial data, and to accurately benchmark against companies in the same industry. In view of the above problems, this invention is proposed to provide a method, apparatus, and equipment for multi-dimensional risk assessment of financial data for oil and gas enterprises that overcomes or at least partially solves the above problems.

[0048] Example 1

[0049] Embodiment 1 of the present invention provides a method for training a multi-dimensional risk assessment model. Referring to Figure 1, the training method may include the following steps:

[0050] Step S11: Obtain the training sample set, wherein each sample in the training sample set includes a different dimension indicator system and its risk preference value; wherein the risk preference value is set based on the experience of financial experts in the oil and gas industry.

[0051] In the specific implementation of this step, the relevant data of the three major financial statements in the historical financial data of oil and gas companies are first obtained; then, based on the screening of general financial indicators and industry-specific financial indicators of various dimensions of oil and gas companies' finances by industry experts; finally, based on the relevant data of the three major financial statements in the historical financial data, and with the general financial indicators and industry-specific financial indicators of various dimensions as the data architecture, a high-dimensional indicator system based on the underlying financial statement data of oil and gas companies' historical financial data is formed.

[0052] The inventors discovered in their practical work that the financial data of the oil and gas industry is easily affected by various factors, exhibiting unique styles and differences. These factors can lead to significant fluctuations in the financial data of the oil and gas industry. Simultaneously, the profitability of the oil and gas industry can be affected by factors such as changes in demand structure and tiered price reductions. These factors can cause fluctuations in the gross profit margin and net profit margin of the oil and gas industry. Furthermore, the oil and gas industry typically invests heavily in R&D, which may affect profit margins in the short term, but in the long run, it may bring new growth opportunities. The inventors further found that existing technologies using general financial data to train convolutional neural networks (CNNs) perform poorly when applied to financial profiling and evaluation in the oil and gas industry, generally exhibiting the problem of excessively low scores for some financial data. To address this issue, this invention improves the accuracy of the model by establishing a financial dataset specifically for the oil and gas sector and performing hierarchical processing on the dataset. Specifically, conventional methods for constructing financial datasets require obtaining large amounts of data from the three major financial statements (balance sheet, income statement, and cash flow statement) and organizing business logic according to conventional financial formulas. To improve the accuracy of this specific industry model, the dataset was designed to discard non-universal data from the three major financial statements. In addition, it incorporated the experience of financial experts in the oil and gas industry to create a financial dataset for the oil and gas industry and set risk preference values ​​specific to the oil and gas industry.

[0053] In this step, industry experts screen out general financial indicators and industry-specific financial indicators for various dimensions of oil and gas companies' finances. Specifically, hierarchical indicators are built based on the collected domain datasets, and industry experts select general financial indicators and industry-specific financial indicators for various dimensions of finances.

[0054] In this embodiment of the invention, general financial indicators refer to upper-level indicator data obtained iteratively from the account transaction amounts and cumulative balance data acquired by the financial system. These mainly include net assets, asset-liability ratio, net profit margin, return on total assets, return on net assets, accounts receivable turnover, inventory turnover, revenue growth rate, and total profit growth rate. Industry-specific financial indicators refer to oil and gas-specific financial indicators provided by oil and gas industry experts, such as technology investment ratio, excess equity guarantee, excess group guarantee plan guarantee, external group guarantee, budget over-management indicators, centralized settlement rate, and authorized network connection rate.

[0055] Furthermore, among general financial indicators, net assets reflect a company's equity capital, which is the net amount after deducting liabilities from assets. It reflects the company's capital structure and financial soundness. The debt-to-equity ratio measures how much of a company's assets are financed through debt, reflecting its financial leverage and debt repayment risk. Net profit margin represents the proportion of net profit to a company's main business revenue, reflecting its profitability and cost control level. Return on assets (ROA) measures a company's ability to generate net profit using its assets, reflecting asset operating efficiency. Return on equity (ROE) reflects the return on equity and is a core indicator for evaluating a company's capital management efficiency. Accounts receivable turnover measures the speed at which a company collects accounts receivable, reflecting its collection efficiency and credit management level. Inventory turnover measures the speed at which a company's inventory turns over, reflecting its inventory management capabilities and market sales performance. Revenue growth rate shows the growth of a company's sales revenue and is an important indicator for measuring a company's market share expansion and market competitiveness. Total profit growth rate reflects the speed of a company's profit growth and is an important indicator for evaluating a company's profitability improvement and development potential. These metrics are widely accepted and used by businesses across various industries because they provide a comprehensive view of a company's financial health.

[0056] Among industry-specific financial indicators, the technology investment ratio reflects a company's level of investment in R&D and technological innovation. For oil and gas companies, R&D is crucial for maintaining a competitive edge and meeting the needs of all parties; a high technology investment ratio signifies strong innovation capabilities and development potential. Excessive equity ratio guarantees refer to guarantees provided by a company to its subsidiaries or affiliates exceeding its equity stake in the guaranteed company. Due to their unique nature, oil and gas companies may be involved in numerous cooperation and subcontracting projects; excessive equity ratio guarantees reflect the company's risk-taking in these collaborations. Guarantees exceeding group limits refer to guarantees provided by oil and gas companies within or to external parties, reflecting the company's credit and risk management capabilities within the supply chain or cooperation network. External group guarantees refer to guarantees provided by oil and gas companies to external parties, which may include guarantees to suppliers, customers, or other partners. This indicator reflects the company's creditworthiness and risk-taking in a broader business environment. Budget over-management indicators refer to the strict budget management and cost control typically involved in oil and gas projects; budget over-management indicators reflect a company's capabilities in budget execution and cost control. Centralized settlement rate refers to the proportion of transactions processed through a centralized settlement system, reflecting the company's efficiency in financial management and capital operation. These distinctive indicators in the embodiments of the present invention help to more comprehensively assess the financial health and operational efficiency of oil and gas companies.

[0057] Step S12: Train the LightGBM model using samples from the training sample set. The different dimension index systems in the samples are passed through the correlation layer of the LightGBM model to obtain the risk preference values ​​corresponding to the different dimension index systems in the samples. The parameters of the LightGBM model are estimated to obtain a multi-dimensional risk assessment model.

[0058] This step uses historical financial data and corresponding risk preference values ​​(thresholds for oil and gas industry-specific indicators) to train the LightGBM model. During training, the model parameters of the LightGBM model are estimated to obtain a multi-dimensional risk assessment model. The output is a score or rating, where the rating is the range within which the score falls. For example, a backtesting accuracy of 85% or higher is considered a successful training result.

[0059] The LightGBM model described in this embodiment of the invention is a gradient boosting framework based on the decision tree algorithm, primarily used for regression and classification problems. The core of LightGBM lies in its histogram algorithm.

[0060] LightGBM, a histogram-based decision tree algorithm, reduces memory consumption and improves the complexity of data segmentation. It discretizes continuous floating-point features into multiple discrete values, constructs a histogram, and then calculates the cumulative statistics of each discrete value in the histogram to find the optimal split point. This method not only reduces memory consumption and computational cost, but also provides a regularization effect due to feature discretization, helping to prevent overfitting.

[0061] Leaf-wise growth strategy: LightGBM employs a depth-constrained leaf-wise growth strategy, splitting the leaf with the highest split gain from all current leaves each time. This strategy can reduce more error and improve model accuracy with the same number of splits. The drawback of leaf-wise growth is that it can potentially lead to overfitting by growing very deep trees. Therefore, LightGBM adds a maximum depth constraint to prevent overfitting.

[0062] Histogram Subtraction Speedup: LightGBM leverages the property that the histogram of a leaf node can be obtained by subtracting the histograms of its parent and sibling nodes, reducing computational cost. This method can obtain the histograms of its sibling leaves at a very low cost after constructing the histogram of a leaf node, thus doubling the speed.

[0063] Categorical Feature Support: LightGBM directly supports categorical features without requiring additional encoding or one-hot unpacking, improving space and time efficiency. It adds decision rules for categorical features to the decision tree algorithm, employing a Many vs Many splitting approach to achieve optimal categorical feature splitting.

[0064] Parallelism Support and Optimization: LightGBM natively supports parallel learning, including feature parallelism and data parallelism. In feature parallelism, different machines search for the optimal split point on different feature sets, and then synchronize the optimal split point. Data parallelism allows different machines to first construct histograms locally, and then perform global merging, searching for the optimal split point on the merged histogram. LightGBM also achieves linear acceleration of parallel computation by using a ensemble communication algorithm.

[0065] In summary, LightGBM provides a fast, low-memory, highly accurate data science tool that supports parallel and large-scale data processing through a series of optimizations.

[0066] The key formulas and parameter definitions involved in the LightGBM model are as follows:

[0067] 1. Learning Rate:

[0068] The learning rate, also known as the shrinkage rate, is a parameter in the gradient boosting algorithm that controls the step size. It determines the proportion of each tree's contribution to the final prediction result. The formula for the learning rate can be expressed as:

[0069]

[0070] Among them, Tree i This is the prediction result for the i-th tree, where n is the total number of trees (number of generations). In this embodiment of the invention, the learning rate is 0.07.

[0071] 2. Number of Estimators / Num Iterations:

[0072] The number of iterations, also known as the number of trees or boosting rounds (num_iterations), refers to the total number of trees built during the boosting process. This parameter directly affects the model's complexity and training time. In this embodiment of the invention, the number of iterations is set to 300.

[0073] 3. Maximum depth of the tree:

[0074] The maximum tree depth (max_depth) limits the maximum depth of the tree model, which can prevent overfitting when the data volume is small. A large tree depth may lead to overfitting, while a small depth may lead to underfitting. In this embodiment of the invention, the maximum tree depth is set to 4.

[0075] 4. Number of Leaf Nodes:

[0076] The `Num_leaves` parameter controls the number of leaf nodes in each decision tree. A larger `Num_leaves` can improve accuracy on the training set, but it also increases the risk of overfitting. In this embodiment of the invention, the number of leaf nodes can be set to the default value.

[0077] 5. Minimum number of samples in a leaf node (Min Data In Leaf):

[0078] This parameter (min_data_in_leaf) represents the minimum number of samples contained in a leaf node and is an important parameter for dealing with overfitting in leaf-wise trees. Setting it to a larger value can avoid generating an overly deep tree, but it may also lead to underfitting. In this embodiment of the invention, the minimum number of samples in the leaf node can be taken as the default value.

[0079] 6. Minimum Hessian Sum of Leaf Nodes:

[0080] This parameter (min_sum_hessian_in_leaf) represents the minimum sum of Hessians at a leaf node, which is the minimum sum of the sample weights at the leaf node, and is used to handle overfitting. In this embodiment of the invention, the minimum leaf node Hessian sum (Min Sum Hessian In Leaf) can take the default value.

[0081] 7. Feature Fraction:

[0082] The feature sampling ratio (feature_fraction) is a floating-point number with a value range of [0, 0, 1, 0], and a default value of 1.0. If it is less than 1.0, LightGBM will randomly select a subset of features for training in each iteration, which can be used to accelerate training and handle overfitting. In this embodiment of the invention, the feature sampling ratio can be set to the default value.

[0083] 8. Bagging Fraction

[0084] The sampling ratio (bagging_fraction) is also a floating-point number, ranging from [0.0, 1.0], with a default value of 1.0. If it is less than 1.0, LightGBM will randomly select a subset of samples for training in each iteration (non-repeated sampling), which can be used to accelerate training and handle overfitting. In this embodiment of the invention, the sampling ratio can be set to the default value.

[0085] The multi-dimensional risk assessment model obtained by the training method provided in this embodiment of the invention has the following beneficial effects when applied:

[0086] (1) Improve the efficiency and accuracy of financial analysis, quickly process massive amounts of historical financial data, and generate accurate and comprehensive indicator values ​​and scores for each dimension in a short period of time to provide timely support for decision-making. For example, it can complete multi-dimensional analysis of the annual financial data of multiple subsidiaries in a single day, which previously took several weeks. Reduce errors and biases caused by manual data processing and ensure the reliability and consistency of analysis results.

[0087] (2) To achieve precise benchmarking and competitive advantage assessment, providing subsidiaries of the oil and gas group with an effective means to accurately benchmark against companies of similar size in the same industry, and clearly understand their own position in the industry. For example, accurately comparing the differences between the company and benchmark companies in key financial indicators such as R&D input-output ratio and cost control efficiency. In-depth analysis of the financial gaps and advantages between the company and benchmark companies provides a basis for formulating targeted development strategies.

[0088] (3) Optimize resource allocation and strategic decision-making, helping oil and gas groups to allocate resources more rationally and improve resource utilization efficiency based on financial analysis results. For example, investing more funds in business areas with competitive advantages, or adjusting poorly performing businesses. Provide strong data support for the group's strategic planning, making decisions more scientific and reasonable.

[0089] (4) Enhance financial risk management capabilities. Through in-depth analysis of financial data, promptly identify potential financial risks and take preventative measures in advance. For example, provide early warnings of potential risks such as cash flow disruptions and debt defaults. This will help build a more robust financial system for the oil and gas group and ensure the company's sustainable development.

[0090] (5) Promote innovation in financial management within the oil and gas industry, introduce advanced technologies and methods, and improve the overall level of financial management and informatization within the industry. Set new benchmarks and models for the oil and gas industry in the field of financial data analysis, thereby promoting the development and progress of the entire industry.

[0091] In another optional embodiment, after obtaining the training sample set, data cleaning is further performed on each sample in the training sample set. This involves removing dirty data or supplementing blank data to achieve the purpose of computer-aided data recognition and processing during the training process.

[0092] Based on the same inventive concept, this embodiment of the invention also provides a multi-dimensional risk assessment model training device. Referring to FIG2, the training device may include: a second acquisition module 21 and a training module 22, and its working principle is as follows:

[0093] The second acquisition module 21 is used to acquire a training sample set, wherein each sample in the training sample set includes a different dimension indicator system and its risk preference value; wherein the risk preference value is set based on the experience of financial experts in the oil and gas industry;

[0094] Training module 22 is used to train the LightGBM model with samples from the training sample set. The different dimension index systems in the samples are passed through the correlation layer of the LightGBM model to obtain the risk preference values ​​corresponding to the different dimension index systems in the samples, so as to perform parameter estimation on the LightGBM model to obtain a multi-dimensional risk assessment model.

[0095] In another optional embodiment, the second acquisition module 21 described above is specifically used for:

[0096] Obtain relevant data from the three major financial statements of oil and gas companies in their historical financial data.

[0097] Based on the selection of general financial indicators and industry-specific financial indicators for various dimensions of oil and gas enterprise finance by industry experts.

[0098] Based on the relevant data from the three major financial statements in historical financial data, and using general financial indicators and industry-specific financial indicators of various dimensions as the data architecture, a high-dimensional indicator system based on the underlying financial statement data is formed for the historical financial data of oil and gas companies.

[0099] The general financial indicators include: net assets, debt-to-equity ratio, net profit margin, return on total assets, return on equity, accounts receivable turnover, inventory turnover, revenue growth rate, and / or total interest growth rate.

[0100] The industry-specific financial indicators include: technology investment ratio, guarantees exceeding equity ratio, guarantees exceeding the group's guarantee plan, guarantees outside the group, budget overrun indicators, and / or centralized settlement rate.

[0101] Based on the same inventive concept, this embodiment of the invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-mentioned multi-dimensional risk assessment model training method.

[0102] Based on the same inventive concept, this embodiment of the invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the above-mentioned multi-dimensional risk assessment model training method.

[0103] The principles by which the above-described apparatus, medium, related equipment, and system in the embodiments of the present invention solve the problem are similar to those of the aforementioned methods. Therefore, their implementation can refer to the implementation of the aforementioned methods, and repeated details will not be repeated.

[0104] Example 2

[0105] Embodiment 2 of the present invention provides a multi-dimensional risk assessment method for financial data of oil and gas enterprises. Referring to Figure 3, the method may include the following steps:

[0106] Step S31: Obtain the current financial data of oil and gas companies.

[0107] Step S32: Based on the relevant data from the three major financial statements in the current financial data, construct a high-dimensional indicator system for the underlying financial statement data. This high-dimensional indicator system includes general financial indicators and industry-specific financial indicators. In practice, this step involves using the current financial data and industry experts to select general and industry-specific financial indicators for various dimensions of oil and gas enterprise finances. Based on the current financial data and using these general and industry-specific financial indicators as the data architecture, a high-dimensional indicator system for the underlying financial statement data is constructed.

[0108] Step S33: Input the high-dimensional indicator system of the underlying report data into the pre-trained multi-dimensional risk assessment model to determine the multi-dimensional risk assessment values ​​of the oil and gas enterprise's financial data. The multi-dimensional risk assessment model described in this embodiment of the invention can be pre-trained according to the multi-dimensional risk assessment model training method in Embodiment 1.

[0109] The method provided in this embodiment of the invention, firstly, improves the quality of financial analysis, reduces errors and omissions caused by manual analysis, and ensures the consistency and reliability of data interpretation. Secondly, it enhances the scientific nature of decision-making, providing strong support for the group's strategic decisions based on accurate financial analysis and benchmarking results, improving the scientific and rational nature of decisions; helping the group to predict market trends and potential risks in advance, and adjust strategies in a timely manner to cope with changes. Thirdly, it promotes competition and development among subsidiaries, enabling subsidiaries to understand their own gaps and advantages compared with the industry peers, clarify the direction for improvement, and stimulate their motivation to improve financial performance and competitiveness; promoting healthy competition among subsidiaries, and jointly improving the financial performance of the entire group. Fourthly, it enhances the industry position, enabling the oil and gas group to be at the forefront of the industry in financial management, enhancing its influence and competitiveness within the industry; providing reference for other companies in the same industry, and promoting the innovation and development of financial management in the entire oil and gas industry. Fifthly, it strengthens risk management, promptly identifying potential financial risks, such as liquidity risks and debt risks, and taking corresponding preventive measures; ensuring the stability and sustainable development of the group's finances.

[0110] Furthermore, in its implementation, this method can pre-divide the contemporaneous financial data of the companies to be benchmarked into blocks, build a large knowledge base for benchmarking question-and-answer models, and then, when engaging with the large model, the model will integrate and analyze the group's own financial data and scores with the contemporaneous data of the benchmark companies, point out the risk points and operational highlights of the financial situation of the inquired subsidiaries, and analyze its own strengths and weaknesses compared with other companies from multiple dimensions, providing guidance and suggestions for each unit to identify gaps, make up for deficiencies, and achieve continuous and stable operation.

[0111] In an optional embodiment, the multi-dimensional risk assessment model is pre-trained using the following method:

[0112] A training sample set is obtained, where each sample includes a different dimension of indicator system and its risk preference value; the risk preference value is set based on the experience of financial experts in the oil and gas industry; specifically, to obtain the training sample set, firstly, relevant data from the three major financial statements of oil and gas companies are obtained from their historical financial data; then, based on the screening by industry experts, general financial indicators and industry-specific financial indicators for each dimension of oil and gas company finances are selected; finally, based on the relevant data from the three major financial statements in the historical financial data, and using the general financial indicators and industry-specific financial indicators for each dimension as the data architecture, a high-dimensional indicator system based on the underlying financial statement data of oil and gas company historical financial data is formed.

[0113] The LightGBM model is trained using samples from the training sample set. The different dimensional indicators in the samples are passed through the correlation layer of the LightGBM model to obtain the risk preference values ​​corresponding to the different dimensional indicators in the samples. The parameters of the LightGBM model are then estimated to obtain a multi-dimensional risk assessment model.

[0114] In another alternative embodiment, general financial metrics may include: net assets, debt-to-equity ratio, net operating profit margin, return on total assets, return on equity, accounts receivable turnover, inventory turnover, revenue growth rate, and / or total interest growth rate.

[0115] Industry-specific financial indicators may include: technology investment ratio, guarantees exceeding equity ratio, guarantees exceeding the group's guarantee plan, guarantees outside the group, budget overrun indicators, and / or centralized settlement rate.

[0116] In another alternative embodiment, after obtaining the training sample set, data cleaning may also be performed on each sample in the training sample set.

[0117] This invention provides a specific example to further illustrate the multi-dimensional risk assessment method for the financial data of oil and gas companies. The example uses the financial data of an oil and gas company for a specific year for rating purposes, as shown in Table 1 below:

[0118] Table 1 Risk Assessment Table of Financial Data of an Oil and Gas Company for a Certain Year

[0119]

[0120]

[0121] In this example, after determining the risk preference values ​​(enterprise scores) of each dimension of the financial data of oil and gas companies through the model, it is possible to further determine the multi-dimensional risk assessment values ​​(total scores) and ratings to replace expert ratings, thereby improving the efficiency and accuracy of financial analysis, achieving precise benchmarking and competitive advantage assessment, optimizing resource allocation and strategic decision-making, further enhancing financial risk management capabilities, and promoting financial management innovation in the oil and gas industry.

[0122] Based on the same inventive concept, this embodiment of the invention also provides a multi-dimensional risk assessment device for financial data of oil and gas enterprises. Referring to FIG4, the device may include: a first acquisition module 41, a construction module 42, and a determination module 43, and its working principle is as follows:

[0123] The first acquisition module 41 is used to acquire the current financial data of oil and gas companies;

[0124] The construction module 42 is used to construct a high-dimensional indicator system for the underlying report data based on the relevant data of the three major financial statements in the current financial data; wherein, the high-dimensional indicator system includes general financial indicators and industry-specific financial indicators;

[0125] The determination module 43 is used to input the high-dimensional indicator system of the underlying report data into the pre-trained multi-dimensional risk assessment model to determine the multi-dimensional risk assessment value of the financial data of oil and gas enterprises.

[0126] Based on the same inventive concept, this embodiment of the invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-mentioned multi-dimensional risk assessment method for financial data of oil and gas enterprises.

[0127] Based on the same inventive concept, this embodiment of the invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the above-mentioned multi-dimensional risk assessment method for financial data of oil and gas enterprises.

[0128] The principles by which the above-described apparatus, medium, related equipment, and system in the embodiments of the present invention solve the problem are similar to those of the aforementioned methods. Therefore, their implementation can refer to the implementation of the aforementioned methods, and repeated details will not be repeated.

[0129] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0130] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or one or more blocks of the block diagrams.

[0131] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0132] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide steps for implementing the functions specified in one or more flowcharts and / or one or more block diagrams.

[0133] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A multi-dimensional risk assessment method for financial data of oil and gas enterprises, characterized in that, include: Obtain the current financial data of oil and gas companies, and based on the relevant data of the three major financial statements in the current financial data, construct a high-dimensional indicator system for the underlying financial data; wherein, the high-dimensional indicator system includes general financial indicators and industry-specific financial indicators; input the high-dimensional indicator system of the underlying financial data into a pre-trained multi-dimensional risk assessment model to determine the multi-dimensional risk assessment value of the financial data of oil and gas companies.

2. The method according to claim 1, characterized in that, The multi-dimensional risk assessment model is pre-trained using the following method: a training sample set is obtained, wherein each sample in the training sample set includes different dimensional indicator systems and their risk preference values; wherein the risk preference values ​​are set based on the experience of financial experts in the oil and gas industry; the LightGBM model is trained using the samples in the training sample set, wherein the different dimensional indicator systems in the samples are passed through the correlation layer of the LightGBM model to obtain the risk preference values ​​corresponding to the different dimensional indicator systems in the samples, so as to perform parameter estimation on the LightGBM model to obtain the multi-dimensional risk assessment model.

3. The method according to claim 2, characterized in that, Obtaining a training sample set includes: acquiring relevant data from the three major financial statements of oil and gas companies in their historical financial data; selecting general financial indicators and industry-specific financial indicators for each dimension of oil and gas companies' finances based on industry experts; and forming a high-dimensional indicator system for the historical financial data of oil and gas companies based on the underlying financial statement data, using the relevant data from the three major financial statements in the historical financial data as the foundation and the general financial indicators and industry-specific financial indicators for each dimension as the data architecture.

4. The method according to claim 3, characterized in that, The general financial indicators include: net assets, debt-to-equity ratio, net operating profit margin, return on total assets, return on net assets, accounts receivable turnover, inventory turnover, revenue growth rate, and / or total interest growth rate; the industry-specific financial indicators include: technology investment ratio, guarantees exceeding equity ratio, guarantees exceeding the group's guarantee plan, guarantees outside the group, budget over-management indicators, and / or centralized settlement rate.

5. The method according to any one of claims 2 to 4, characterized in that, After obtaining the training sample set, the process also includes data cleaning for each sample in the training sample set.

6. A method for training a multi-dimensional risk assessment model, characterized in that, include: A training sample set is obtained, wherein each sample in the training sample set includes different dimension indicator systems and their risk preference values; wherein the risk preference values ​​are set based on the experience of financial experts in the oil and gas industry; the LightGBM model is trained using the samples in the training sample set, wherein the different dimension indicator systems in the samples are passed through the correlation layer of the LightGBM model to obtain the risk preference values ​​corresponding to the different dimension indicator systems in the samples, so as to perform parameter estimation on the LightGBM model to obtain a multi-dimensional risk assessment model.

7. A multi-dimensional risk assessment device for financial data of oil and gas enterprises, characterized in that, include: The first acquisition module is used to acquire the current financial data of oil and gas companies; The construction module is used to construct a high-dimensional indicator system for the underlying financial data based on the relevant data of the three major financial statements in the current financial data; wherein, the high-dimensional indicator system includes general financial indicators and industry-specific financial indicators; the determination module is used to input the high-dimensional indicator system of the underlying financial data into a pre-trained multi-dimensional risk assessment model to determine the multi-dimensional risk assessment value of the financial data of oil and gas enterprises.

8. A multi-dimensional risk assessment model training device, characterized in that, include: The second acquisition module is used to acquire a training sample set, wherein each sample in the training sample set includes a different dimension indicator system and its risk preference value; wherein the risk preference value is set based on the experience of financial experts in the oil and gas industry; The training module is used to train the LightGBM model with samples from the training sample set. The different dimension index systems in the samples are passed through the correlation layer of the LightGBM model to obtain the risk preference values ​​corresponding to the different dimension index systems in the samples, so as to perform parameter estimation on the LightGBM model to obtain a multi-dimensional risk assessment model.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the multi-dimensional risk assessment method for financial data of oil and gas enterprises as described in any one of claims 1 to 5, or the multi-dimensional risk assessment model training method as described in claim 6.

10. A computer 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 program, it implements the multi-dimensional risk assessment method for financial data of oil and gas enterprises as described in any one of claims 1 to 5, or implements the multi-dimensional risk assessment model training method as described in claim 6.

Images