Machine learning-based financial feasibility evaluation system and method for green financial project

Abstract

The invention relates to the technical field of green financial project financial feasibility evaluation systems, in particular to a green financial project financial feasibility evaluation system and method based on machine learning, and the system comprises a financial evaluation module, a market evaluation module and an environment evaluation module. The financial evaluation module comprises an analysis evaluation unit, a prediction analysis unit and a model simulation unit; the analysis and evaluation unit is used for collecting and sorting data of items and clients and processing the data; the prediction analysis unit predicts financial indexes under different market scenes by utilizing an optimization algorithm according to the customer data and the market data; the market assessment module provides key external environment information for financial assessment by deeply analyzing market trend and policy changes. Meanwhile, the environment evaluation module quantifies the environment influence of the project, and provides important environment indexes for market evaluation and financial evaluation. The complementary mechanism ensures that all key factors influencing the feasibility of the project are considered in the evaluation process.

Description

Technical Field

[0001] The present invention relates to the technical field of green finance project financial feasibility assessment systems, and in particular to a green finance project financial feasibility assessment system and method based on machine learning. Background Art

[0002] In recent years, with the increasing global attention to environmental issues, green finance has become an important tool to promote sustainable development. However, the evaluation of green finance projects has always been a complex and challenging issue. Traditional project evaluation methods mainly focus on financial indicators, such as net present value (NPV) and internal rate of return (IRR), and often ignore the key factor of environmental benefits. This evaluation method is difficult to fully reflect the true value of green projects, resulting in some projects with significant environmental benefits but low short-term financial returns being wrongly rejected.

[0003] With the development of assessment technology, some institutions have begun to try to incorporate environmental factors into the assessment system. For example, some assessment methods use concepts such as carbon credits to quantify environmental benefits. However, these methods are usually static and difficult to adapt to the rapidly changing market and policy environment. In addition, they often use environmental benefits as an independent assessment dimension rather than organically combining them with financial assessment. This fragmented assessment method makes it difficult to provide decision makers with a comprehensive and integrated project value judgment.

[0004] Another significant problem is that most existing evaluation methods rely on human experience and subjective judgment. This not only increases the cost and time of the evaluation, but also introduces human bias, affecting the objectivity and accuracy of the evaluation. Especially when facing emerging green technology projects, traditional rules of thumb are no longer applicable, resulting in a significant reduction in the reliability of the evaluation results.

[0005] In addition, existing evaluation methods are usually unable to effectively deal with uncertainty. Green finance projects often face complex market environments and policy risks, and traditional deterministic evaluation methods are difficult to accurately capture the impact of these uncertainties on project value. Although some methods introduce scenario analysis, they usually only consider a limited number of scenarios and are unable to fully reflect the complexity of the real world.

[0006] In view of the above problems, a new method is urgently needed to comprehensively, dynamically and objectively evaluate green finance projects. The ideal evaluation system should be able to: 1) organically combine environmental benefits with financial benefits; 2) adapt to the rapidly changing market and policy environment; 3) minimize the impact of human bias; 4) effectively deal with uncertainty; 5) provide decision makers with comprehensive and intuitive project value judgments. Summary of the invention

[0007] The present invention is an innovative solution to these urgent needs. By introducing machine learning technology, the present invention proposes a green finance project financial feasibility assessment system and method based on machine learning. The system integrates three modules: financial assessment, market assessment, and environmental assessment, and realizes the monetization of environmental benefits and dynamic optimization of financial indicators in a data-driven manner. The core of the system lies in its adaptability and learning ability. It can continuously learn from new data and continuously optimize the assessment model, thereby providing more accurate and comprehensive project assessment results.

[0008] The present invention proposes a green finance project financial feasibility assessment system based on machine learning, comprising a financial assessment module, a market assessment module and an environmental assessment module; the financial assessment module comprises an analysis and assessment unit, a prediction and analysis unit and a model simulation unit; the analysis and assessment unit is used to collect and organize project and customer data and process the data; the prediction and analysis unit predicts financial indicators under different market scenarios using an optimization algorithm based on customer data and market data; the model simulation unit provides risk scenario simulation and cash flow simulation of recovery income under different scenarios; the market assessment module comprises a data analysis unit and a prediction and analysis unit, the data analysis unit collects market data and industry data and screens them according to the classification of green finance projects; the prediction and analysis unit predicts changes in technical parameters of green projects under different market prices or environmental policy scenarios under future scenarios based on the screened data; the environmental assessment module comprises a data analysis unit, a model simulation unit and an environmental assessment unit, the data analysis unit collects and analyzes environmental data of the project location; the model simulation unit performs numerical modeling on the environmental data; the environmental assessment unit judges the acceptability of environmental benefits by comparing the simulated environmental benefit indicators with the set environmental benefit indicator threshold.

[0009] The method of the financial feasibility assessment system of green finance projects based on machine learning includes the following steps: S1. Collecting market data, project data and environmental data of green finance projects, and classifying customers according to the market data and project data to achieve customer classification; S2. The prediction and analysis unit uses an optimization algorithm to predict the changes of different financial indicators of customers under market conditions; S3. The model simulation unit establishes different risk scenarios for the changes of different financial indicators of customers, and calculates the net present value (NPV) of the project through the recovery income cash flow of numerical simulation; S4. The model simulation unit numerically models the changes of different technical indicators of the project under environmental changes, and finally obtains the technical recovery indicator cash flow; S5. The environmental evaluation unit evaluates whether different environmental benefit indicators meet the predetermined threshold.

[0010] The beneficial technical effects brought by the technical solution of the present invention are:

[0011] The green finance project financial feasibility assessment system based on machine learning proposed in the present invention has brought a series of significant beneficial effects to the field of green finance project assessment by cleverly integrating the three modules of financial assessment, market assessment and environmental assessment.

[0012] First, the present invention effectively solves the contradiction between the separation of financial benefits and environmental benefits in traditional evaluation methods. In the financial evaluation module, an optimization algorithm that takes environmental factors into consideration is innovatively introduced, so that environmental benefits can be accurately quantified and integrated into financial evaluation. This integration not only overcomes the problem of insufficient value evaluation of green projects by traditional methods, but also provides decision makers with a more comprehensive and accurate judgment of project value.

[0013] Secondly, the three major evaluation modules show strong complementarity, which significantly improves the comprehensiveness and accuracy of the evaluation. The market evaluation module provides key external environmental information for financial evaluation by deeply analyzing market trends and policy changes. At the same time, the environmental evaluation module quantifies the environmental impact of the project and provides important environmental indicators for market evaluation and financial evaluation. This complementary mechanism ensures that the evaluation process takes into account all key factors affecting the feasibility of the project.

[0014] Furthermore, the algorithm design of the present invention embodies a strong synergistic effect. In the financial assessment module, the forecasting analysis unit and the model simulation unit work together to not only predict financial indicators but also simulate various risk scenarios. This synergistic mechanism significantly improves the reliability of the assessment by considering diversity. At the same time, the data sharing mechanism between the market assessment module and the environmental assessment module further strengthens the synergistic effect, ensures that each module uses a consistent data basis, and improves the consistency and accuracy of the overall assessment.

[0015] In addition, the algorithm design of the present invention also reflects the multiple superposition effect. In the environmental assessment module, the calculation of the cash flow of the environmental benefit indicator multiplies and accumulates the environmental benefit and its monetization coefficient, realizing the superposition of environmental benefits in the time dimension. At the same time, when calculating the final net present value, these environmental benefits are superimposed with traditional financial indicators to form a comprehensive indicator that fully reflects the value of the project. This multiple superposition not only improves the accuracy of the assessment, but also makes the assessment results more comprehensively reflect the overall value of the project.

[0016] Finally, the machine learning algorithm used in the present invention performs well in dealing with complexity and uncertainty. The regularization term in the optimization algorithm effectively prevents overfitting and improves the generalization ability of the model when facing new data. At the same time, the model evaluation and correction mechanism enables the system to continuously learn from new data, dynamically adjust parameters, and adapt to changes in the market and environment. This adaptive ability greatly enhances the applicability of the evaluation system in complex and changeable real-world environments.

[0017] In summary, the present invention effectively solves many contradictions in traditional evaluation methods through innovative algorithm design and module integration, and realizes the organic unity of financial, market and environmental factors. Its complementary, synergistic and superimposed effects significantly improve the comprehensiveness, accuracy and reliability of the evaluation. At the same time, the application of machine learning algorithms enables the system to have strong adaptive capabilities and can effectively cope with complex and changing real environments. These advantages make the present invention have great practical value in the field of green financial project evaluation, and it is expected to promote the healthy development of green finance and make important contributions to global sustainable development. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is the overall logic block diagram of the system of the present invention.

[0019] Figure 2 It is a logic block diagram of the financial evaluation module of the present invention.

[0020] Figure 3 It is a logic block diagram of the market evaluation module of the present invention.

[0021] Figure 4 It is a logic block diagram of the environmental assessment module of the present invention.

[0022] Figure 5 A logic block diagram of the model evaluation modification for the present invention.

[0023] Figure 6 A logical block diagram for generating a project financial analysis report according to the present invention. DETAILED DESCRIPTION

[0024] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects thereof are described in detail below in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0025] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0026] See also Figure 1-6 The present invention provides a green finance project financial feasibility assessment system based on machine learning. The system mainly consists of three core modules: financial assessment module 1, market assessment module 2 and environmental assessment module 3. The three modules work together to complete a comprehensive assessment of green finance projects.

[0027] First, the financial evaluation module 1 is the core of the whole system, which includes three important functional units: analysis and evaluation unit 11, forecast analysis unit 12 and model simulation unit 13. The main task of the analysis and evaluation unit 11 is to collect and organize relevant data of the project and the customer. For example, for a solar power generation project, the analysis and evaluation unit 11 will collect financial data such as the project's investment scale, expected power generation, operating costs, and other information, as well as the customer's credit status, past project experience, and other information.

[0028] Then, the forecasting and analysis unit 12 will use these data, combined with the market data provided by the market assessment module 2, to use an optimization algorithm to predict financial indicators under different market scenarios. The optimization algorithm used here is a method based on machine learning, which obtains the optimal prediction function by minimizing the weighted sum of the prediction error and the model complexity. Specifically, the objective function of the algorithm is:

[0029] F = argmin(∑(y i -f(x i )) 2 +λ||f|| 2 )

[0030] In this formula, F represents the optimal prediction function we are looking for, y i is the actual financial indicator value (such as the internal rate of return of the project), x i is the influencing factor (such as electricity price, equipment cost, etc.), f(x i ) is the predicted value, and λ is a parameter used to balance the prediction accuracy and model complexity. Through this algorithm, the prediction analysis unit 12 can obtain a model that can accurately predict financial indicators without overfitting.

[0031] Next, the model simulation unit 13 will perform risk scenario simulation and cash flow simulation based on the results of the forecast analysis unit 12. For example, it will simulate the project cash flow under different scenarios such as a 10% drop in electricity prices and a 5% increase in equipment efficiency. These simulation results provide an important basis for subsequent decision-making.

[0032] Secondly, the market assessment module 2 includes a data analysis unit 21 and a forecast analysis unit 22. The data analysis unit 21 is responsible for collecting and screening market data. In green finance projects, these data include energy price trends, environmental policy changes, technology development trends, etc. Based on these data, the forecast analysis unit 22 predicts the impact of future market environment changes on the project. For example, it predicts the rate of improvement of photovoltaic power generation technology efficiency in the next five years, thereby evaluating the long-term competitiveness of the project.

[0033] Thirdly, the environmental assessment module 3 is a major feature of the system, which includes a data analysis unit 31, a model simulation unit 32 and an environmental assessment unit 33. The data analysis unit 31 mainly collects environmental data of the project location, such as air quality, water quality, biodiversity and other indicators. The model simulation unit 32 performs numerical modeling on these data to predict the potential impact of the project on the environment. Finally, the environmental assessment unit 33 compares the simulation results with the preset environmental benefit index threshold to determine whether the environmental benefits of the project are acceptable.

[0034] Suppose we are evaluating a wind power project. Environmental Assessment Module 3 will first collect data on bird migration routes, noise levels, etc. at the project site. Then, numerical models will be used to predict the impact of wind turbines on bird habitats and local noise levels. Finally, these predictions will be compared with preset thresholds (such as a 5% reduction in the number of birds and a 3 dB increase in noise) to determine whether the environmental impact of the project is within an acceptable range.

[0035] Through the collaborative work of these three modules, this system can comprehensively evaluate the financial feasibility, market prospects and environmental benefits of green finance projects, providing a scientific and objective basis for investment decisions.

[0036] In a preferred embodiment of the present invention, data sharing is implemented between the market assessment module 2 and the environmental assessment module 3. This data sharing mechanism greatly improves the efficiency of the system and the accuracy of the assessment. Specifically, the shared data set DS includes data that exists in both the market assessment module data set M and the environmental assessment module data set E.

[0037] This data sharing method can be expressed by a mathematical expression: DS = {d | d ∈ M | E}. Here, d represents a specific data item, M represents the entire data set of the market assessment module 2, and E represents the entire data set of the environmental assessment module 3. By taking the intersection of the two data sets, we get the shared data set DS.

[0038] Through this data sharing mechanism, the system can analyze the feasibility of the project more comprehensively. For example, if the supply of biomass raw materials is insufficient, the market assessment module 2 will predict rising costs, while the environmental assessment module 3 will predict declining environmental benefits. This multi-angle analysis can help decision makers better understand the potential risks and opportunities of the project.

[0039] In addition, data sharing can also reduce duplication of work and improve system efficiency. In actual operation, we can set up a central database to store the shared data set DS, and the two modules can access this database at any time, thus realizing real-time data sharing and updating.

[0040] Another important feature of the present invention is that the market assessment module 2 uses an efficient screening algorithm for matching market data with green finance projects. The core idea of ​​this screening algorithm is to calculate the weighted similarity between project features and market data features.

[0041] Specifically, the screening algorithm uses the following formula: S(p,m)=∑(w i ·sin(p i ,m i ))

[0042] In this formula, S(p,m) represents the overall matching score between project p and market data m. i is the weight of the i-th feature, reflecting the importance of this feature in the matching process. sin(pi,m i ) is a similarity function used to calculate the i-th feature p of the item i and the i-th feature m of the market data i The degree of similarity between them.

[0043] An important advantage of this screening algorithm is its flexibility. By adjusting the weights and similarity functions, we can easily adapt to different types of green finance projects. For example, for wind power projects, we will add the feature of average wind speed and give it a higher weight.

[0044] In addition, this algorithm can be easily combined with machine learning methods. For example, we can use historical data to train a model to automatically learn the optimal weight distribution. In this way, as the system processes more and more items, the matching accuracy will continue to improve.

[0045] Through this screening algorithm, the market evaluation module 2 can quickly find the market data most relevant to the current project, thereby providing more accurate market evaluation results. This not only improves the efficiency of the evaluation, but also enhances the reliability of the evaluation results.

[0046] In the implementation process of the present invention, the forecasting and analysis unit 12 and the model simulation unit 13 work together to jointly evaluate the impact of market environment changes and recovery income risks on the project cash flow. This collaborative evaluation method is a major feature of the system, which can more comprehensively and accurately reflect the financial feasibility of the project.

[0047] Specifically, the evaluation method uses the Net Present Value (NPV) calculation formula:

[0048]

[0049] Among them, CF is the net present value of cash flow, R t is the profit at time t, Ct is the cost at time t, r is the discount rate, and t is time (usually in years). This formula takes into account the time value of money and discounts future cash flows to current values.

[0050] Let's use a concrete example to illustrate how this assessment method works. Suppose we are assessing a wind power project with a project life of 20 years.

[0051] First, the forecasting and analysis unit 12 will predict key factors such as changes in electricity prices and equipment maintenance costs in the next 20 years based on historical data and market trends. For example, it predicts that electricity prices will increase by 2% each year and maintenance costs will increase by 1.5% each year.

[0052] Then, the model simulation unit 13 will use these forecast results to simulate the project cash flow under different market environments. For example, it will simulate the following three scenarios: Baseline scenario: Development in accordance with the forecast trend; Optimistic scenario: The electricity price rises faster than expected, reaching 3% per year; Pessimistic scenario: The electricity price growth stagnates, and the maintenance cost rises faster to 2.5% per year;

[0053] For each scenario, the model calculates the net present value of cash flows over 20 years. Assuming we choose 8% as the discount rate (this rate is usually determined based on industry standards and project risk level), then for the baseline scenario, the cash flows in the first year are calculated as follows:

[0054] Revenue R1 = 10 million yuan (assumed value);

[0055] Cost C1 = 6 million yuan (assumed value);

[0056] CF1=(10 million 6 million)*(1+0.08)^1≈3.7037 million yuan

[0057] Similarly, we can calculate the discounted cash flow for each year over the 20 years and then sum them up to get the total net present value.

[0058] By comparing the net present value under different scenarios, we can assess the sensitivity of the project to changes in the market environment. For example, if the net present value under the pessimistic scenario is still positive, then we can believe that the project has strong risk resistance.

[0059] In addition, this method can also help us calculate important financial indicators such as the internal rate of return (IRR) and payback period of the project. For example, the internal rate of return is the discount rate r that makes the net present value equal to zero. By adjusting the value of r until CF equals zero, we can find the IRR of the project.

[0060] An important advantage of this evaluation method is that it takes into account time factors and risk factors. This is particularly important in green finance projects, as such projects usually have a long investment cycle and high uncertainty. By simulating different market environments and risk scenarios in detail, we can provide decision makers with more comprehensive and reliable financial evaluation results.

[0061] In a preferred embodiment of the present invention, the environmental data required by the market assessment module 2 is provided by the environmental assessment module 3. This data sharing mechanism not only improves the efficiency of the system, but also ensures the consistency of the environmental data used by each module. However, considering the sensitivity and importance of environmental data, this system adopts encryption measures during data transmission.

[0062] Specifically, data transmission uses a simple but effective encryption method, namely XOR encryption. The encryption formula is:

[0063]

[0064] In this formula, E(D) represents the encrypted data, D is the original data, and K is the key. represents XOR operation. An important feature of XOR operation is that if the same XOR operation is performed on the encrypted data again, the original data can be obtained, that is, This makes the decryption process very simple.

[0065] The advantages of this encryption method are that it has fast calculation speed, simple implementation, and the same algorithm is used for encryption and decryption. This is a good choice for the green finance project evaluation system that needs to process large amounts of data in real time.

[0066] In practical applications, we would use longer data blocks and keys, such as 256 bits or more. In addition, to further improve security, we can change the keys regularly or use different keys for different types of data.

[0067] Through this encryption mechanism, even if the data is intercepted during transmission, the original data cannot be deciphered without the key. This ensures the security of environmental data transmission between modules and prevents data leakage or tampering.

[0068] It is worth noting that this encryption method is mainly used for internal data transmission between modules. For data interaction between the system and the outside, we need to adopt more complex encryption algorithms, such as AES or RSA, to meet more severe security challenges.

[0069] In general, this data encryption mechanism not only ensures the security of data, but also does not significantly increase the computing burden of the system. It is a solution that takes into account both efficiency and security.

[0070] The method for assessing the financial feasibility of green finance projects based on machine learning proposed in this paper is a systematic, multi-step process. This method makes full use of machine learning technology to comprehensively assess the financial feasibility, market prospects and environmental benefits of green finance projects. Let us explain each step in detail below:

[0071] First, step S1 is the basis of the entire assessment process. In this step, the system will comprehensively collect various data related to green finance projects. These data mainly include three categories: market data, project data and environmental data.

[0072] Market data includes energy price trends, relevant policy changes, industry development forecasts, etc. For example, for a photovoltaic power generation project, we need to collect information such as electricity price trends, photovoltaic equipment cost changes, and national subsidy policies.

[0073] Project data refers to information related to specific projects, such as project scale, technical parameters, expected life, investment cost, etc. For example, the installed capacity, wind turbine type, and expected annual power generation of a wind power project.

[0074] Environmental data mainly involve the environmental conditions of the project site and the environmental impacts brought about by the project, such as air quality, water resources, ecosystem diversity, etc.

[0075] Based on the data collected, the system will classify customers according to market data and project data. This classification can help us better understand the characteristics and risks of different types of projects. For example, we will classify according to factors such as project size (large, medium, small), technology maturity (mature technology, emerging technology), and geographical location (urban, suburban, remote areas).

[0076] Secondly, in step S2, the prediction analysis unit 12 uses an optimization algorithm to predict the changes in the financial indicators of the customer under different market conditions. The optimization algorithm here is usually based on machine learning methods, such as support vector machines (SVM), random forests or neural networks. These algorithms can learn patterns in historical data to make more accurate predictions of future financial indicators.

[0077] Next, in step S3, the model simulation unit 13 will establish different risk scenarios based on the prediction results of S2. These scenarios include the best case, the worst case, and several intermediate cases. For each scenario, the system will simulate the cash flow of the project and calculate the net present value (NPV).

[0078] For example, let's say we are evaluating a geothermal energy project. We simulate the following scenarios: Baseline scenario: current trends continue; Optimistic scenario: technological advances lead to a 20% drop in drilling costs; Pessimistic scenario: geothermal resources are not as good as expected, and power generation is 15% lower than expected;

[0079] For each scenario, we calculate the NPV of the project. This allows us to understand the financial performance of the project under different circumstances and assess the project's risk tolerance.

[0080] Then, in step S4, the model simulation unit 13 numerically models the changes in the technical indicators of the project under environmental changes. The purpose of this step is to evaluate the impact of environmental factors on the technical performance of the project. For example, for a solar project, we will simulate how the power generation will change when the sunshine time is reduced due to climate change. Finally, we will get a cash flow model that reflects the technical recovery indicators.

[0081] Finally, in step S5, the environmental assessment unit 33 will evaluate whether the environmental benefit indicators of the project meet the predetermined thresholds. These indicators include carbon emission reduction, air quality improvement, water resource conservation, etc. For example, for a waste incineration power generation project, we will set the following thresholds: carbon emission reduction: at least 100,000 tons per year; waste treatment: at least 1,000 tons per day; dioxin emissions: no more than 0.1ng TEQ / m 3 ;

[0082] Only when all environmental benefit indicators meet or exceed predetermined thresholds will the project be considered environmentally acceptable.

[0083] Through these five steps, this method can comprehensively evaluate the financial feasibility, risk resistance and environmental benefits of green finance projects. This multi-dimensional evaluation method can not only help investors make more informed decisions, but also promote more support for projects that are truly beneficial to the environment, thereby promoting the healthy development of green finance.

[0084] In a preferred embodiment of the present invention, step S1 is further divided into five sub-steps to ensure the comprehensiveness and accuracy of data collection. The specific implementation methods of these sub-steps are as follows:

[0085] First, in step S1.1, the market collection unit will collect market data and industry data regularly. Regularly here means daily, weekly or monthly, and the specific frequency depends on the type and importance of the data. For example, electricity prices need to be updated every day, while it is sufficient to update industry policy changes once a month. The collected data includes but is not limited to: energy price trends, relevant policy changes, technology development trends, market supply and demand conditions, etc.

[0086] Secondly, in step S1.2, the market collection unit will classify the green finance projects. This classification can be based on multiple dimensions, such as project type (solar energy, wind energy, biomass energy, etc.), project scale (large, medium, small), technology maturity (mature technology, emerging technology), etc. For example, a classification system is as follows: Renewable energy generation: a. Solar power generation (large ground power station / distributed roof); b. Wind power generation (onshore / offshore); c. Biomass power generation; d. Geothermal power generation; Energy efficiency improvement: a. Industrial energy-saving transformation; b. Building energy saving; c. Transportation energy saving; Environmental governance: a. Water pollution control; b. Air pollution control; c. Solid waste treatment;

[0087] Then, in step S1.3, the system matches the client’s project data based on the classification of green finance project types. The purpose of this step is to find the most relevant market and industry data for each project. For example, if the client’s project is an offshore wind power project, the system will match offshore wind power-related data, such as offshore wind speed data, offshore wind turbine cost trends, and offshore power transmission technology developments.

[0088] Next, in step S1.4, the system will collect local environmental data. These data come from multiple sources, such as monitoring data from local environmental protection departments, satellite remote sensing data, field survey data, etc. For different types of projects, the environmental data that needs to be collected will also be different. For example, for a solar energy project, we need to collect local sunshine intensity, atmospheric transparency and other data; while for a biomass energy project, we are more concerned with local biomass resource reserves, air quality and other data.

[0089] Finally, in step S1.5, the environmental assessment module 3 will screen, organize and clean the collected environmental data. This step is crucial to ensure the accuracy of subsequent analysis. In particular, data cleaning adopts a statistically based method, whose formula is:

[0090] D clean ={d|d∈D,d-μ≤3σ}

[0091] In this formula, D clean is the cleaned data set, D is the original data set, μ is the mean of the data, and σ is the standard deviation of the data. This method is based on the three sigma principle, that is, in a normal distribution, 99.7% of the data falls within three standard deviations of the mean. With this method, we can effectively remove outliers and improve data quality.

[0092] The advantage of this data cleaning method is that it can automatically adapt to data of different scales without manually setting fixed thresholds. At the same time, it also retains the normal fluctuations of the data and only removes the real outliers. This is especially important for environmental data, because environmental data often has large natural fluctuations.

[0093] Through these five detailed sub-steps, we can ensure that we obtain high-quality and comprehensive data, laying a solid foundation for subsequent analysis and evaluation. This not only improves the accuracy of the evaluation, but also enhances the reliability and adaptability of the entire system.

[0094] In a preferred embodiment of the present invention, step S2 is further refined into five sub-steps to achieve a comprehensive simulation of market scenarios and accurate prediction of financial indicators. The specific implementation methods of these sub-steps are as follows:

[0095] First, in step S2.1, the forecasting and analysis unit 12 provides several market scenario conditions. These scenario conditions include two key components: market scenario parameters and scenario condition probabilities. Market scenario parameters mainly include market electricity prices, market coal prices, market natural gas prices, and market environmental protection policies. These parameters are selected based on their significant impact on the financial feasibility of green finance projects.

[0096] Secondly, in step S2.2, the forecasting and analysis unit 12 will randomly generate several market scenario conditions according to the scenario parameters. The purpose of this step is to increase the diversity of simulation and cover the market conditions more comprehensively. For example, the system will randomly generate 100 slightly changed scenarios based on the baseline scenario, and the parameter values ​​of these scenarios will fluctuate within a certain range.

[0097] Then, in step S2.3, the system will obtain technical benefits under several market scenarios from the model simulation unit 13 according to the generated market scenario conditions. The technical benefits here refer to the economic benefits that the project is expected to generate under given market conditions. For example, for a wind power generation project, its annual power generation income under different electricity price scenarios is a key technical benefit indicator.

[0098] Next, in step S2.4, the forecasting and analysis unit 12 uses an optimization algorithm to optimize the technical benefits based on the generated market scenario conditions and the obtained technical benefits, thereby obtaining financial indicators under different market scenarios. The core of this optimization process is to find a solution that can maximize the expected benefits.

[0099] Finally, in step S2.5, the model simulation unit 13 receives these optimized financial indicators. The formula used by the optimization algorithm is:

[0100] FI=argmax(∑(P i·T i ))

[0101] Among them, FI is the optimized financial indicator, P i is the probability of the ith market scenario, T i is the technical benefit in the i-th market scenario. This formula is essentially calculating the expected benefit and finding the combination of financial indicators that maximizes the expected benefit.

[0102] Through this method, we can obtain a forecast of financial indicators that takes into account multiple factors and is probability-weighted. This forecast is more comprehensive and reliable than a single scenario forecast and can provide better support for project decision-making.

[0103] In another preferred embodiment of the present invention, step S3 is further refined into four sub-steps to achieve a comprehensive assessment of project risks and an accurate calculation of the net present value. The specific implementation methods of these sub-steps are as follows:

[0104] First, in step S3.1, the model simulation unit 13 receives financial indicators under several different market scenarios from the forecasting and analysis unit 12. At the same time, it also sets risk scenario parameters and risk scenario conditional probabilities. Risk scenario parameters include technical risks (such as equipment failure rate), policy risks (such as changes in subsidy policies), natural risks (such as extreme weather events), etc. Each risk scenario will also be assigned a conditional probability, indicating the likelihood of the risk occurring.

[0105] Secondly, in step S3.2, the model simulation unit 13 will establish several risk scenarios for the financial indicators under different market scenarios received previously, based on the risk scenario parameters and risk scenario conditional probabilities just set. The purpose of this step is to combine market scenarios and risk scenarios to form a more comprehensive scenario matrix.

[0106] For example, if we have 3 market scenarios and 4 risk scenarios, we will get 3*4=12 combination scenarios. Each combination scenario has its corresponding probability, which is the product of the market scenario probability and the risk scenario probability.

[0107] Then, in step S3.3, the model simulation unit 13 calculates the project recovery income cash flow NPV0 based on these combined scenarios. NPV0 represents the net present value that the project is expected to generate after considering various market conditions and risk factors. This calculation process takes into account factors such as income, cost, and risk impact under each scenario.

[0108] Finally, in step S3.4, the model simulation unit 13 simulates these risk scenarios and NPV0 to generate several NPVs of recovery income cash flow. The NPV calculation formula used here is:

[0109]

[0110] In this formula, CFt represents the cash flow at time t, r is the discount rate, and I0 is the initial investment amount. This formula takes into account the time value of money and converts future cash flows into current values.

[0111] Let's illustrate this process with a concrete example:

[0112] Suppose we are evaluating a solar power generation project with an initial investment of RMB 100 million and an expected life of 20 years. We choose a combination scenario: the market scenario is that the electricity price remains stable, and the risk scenario is that the equipment wears out early and requires additional maintenance. In this scenario, the cash flow forecast of the project is as follows:

[0113] Year 15: Annual net cash flow is RMB 8 million;

[0114] Year 610: Annual net cash flow is RMB 7.5 million (taking into account the aging of equipment);

[0115] Year 1120: Annual net cash flow is RMB 7 million (taking into account further aging of equipment);

[0116] We assume the discount rate r is 8% (this rate is usually determined based on factors such as industry standards, inflation rate and project risk).

[0117] Applying the NPV formula:

[0118] NPV=(800*(1(1+0.08) 5 ) / 0.08) / (1+0.08) 5 +(750*(1(1+0.08) 5 ) / 0.08) / (1+0.08) 10

[0119] +(700*(1(1+0.08) 10 ) / 0.08) / (1+0.08) 10 10000≈2873.97+1872.68+2188.5310000≈3064.82 (10,000 yuan)

[0120] In this particular scenario, the NPV of the project is $30,648,200, which means that the project is not feasible in this case.

[0121] However, this is only one of many scenarios. The model simulation unit 13 calculates the NPV under all scenarios and calculates the expected NPV based on the probability of each scenario. If the expected NPV is positive and the risk is within an acceptable range, the project is considered financially feasible.

[0122] The advantage of this approach is that it takes into account market changes and various risk factors, and can provide decision makers with more comprehensive and reliable information. By simulating a large number of scenarios, we can better understand the potential returns and risks of the project and make more informed investment decisions.

[0123] In another preferred embodiment of the present invention, step S4 is further refined into three sub-steps, aiming to comprehensively evaluate the environmental benefits and economic value of the project. The specific implementation methods of these sub-steps are as follows:

[0124] First, in step S4.1, the model simulation unit 13 receives several technical indicators under different market scenarios. These technical indicators include power generation, energy conversion efficiency, pollutant emission reduction, etc., which will change with the changes in market scenarios. For example, in a market scenario with higher electricity prices, more efficient but more expensive equipment will be used to improve power generation and energy conversion efficiency.

[0125] Secondly, in step S4.2, the model simulation unit 13 receives these different technical indicators and performs modeling to obtain environmental economic benefit indicators under several different scenarios. The key to this step is to convert technical indicators into quantifiable environmental benefits. For example, for a wind power project, we will focus on the following environmental economic benefit indicators:

[0126] Carbon reduction: The amount of carbon dioxide emissions reduced by replacing fossil fuel electricity generation;

[0127] Improved air quality: reduced emissions of pollutants such as sulfur dioxide and nitrogen oxides;

[0128] Water conservation: the amount of water conserved compared to traditional thermal power plants;

[0129] Land use efficiency: the amount of clean energy produced per unit area of ​​land;

[0130] The model will calculate the specific values ​​of these environmental and economic benefit indicators in each scenario based on different technical indicators.

[0131] Finally, in step S4.3, the model simulation unit 13 simulates the corresponding environmental benefit indicator cash flow according to the environmental economic benefit indicators under these different scenarios. The core of this step is to monetize environmental benefits, that is, to convert environmental benefits into quantifiable economic value. The calculation formula of environmental benefit indicator cash flow is:

[0132] ECFV=∑EB t ·P t

[0133] In this formula, ECFV represents the total value of environmental benefit indicator cash flow, EB t represents the environmental benefit at time t (such as carbon emission reduction), and P represents the monetization coefficient of the environmental benefit at time t (such as carbon price).

[0134] Let's illustrate this process with a concrete example:

[0135] Suppose we are evaluating a wind power project with an annual power generation capacity of 100 million kWh and a project life of 20 years. We are mainly interested in its carbon reduction benefits.

[0136] Step S4.1: Under the baseline market scenario, technical indicators show that wind power can displace 0.8 kg of CO2 emissions per kilowatt-hour (this value depends on the local power mix).

[0137] Step S4.2: Based on this, we can calculate the annual carbon emission reduction of the project:

[0138] 100 million kWh*0.8 kg / kWh=80,000 tons of CO2

[0139] Step S4.3: Assume that the carbon price (i.e., the environmental benefit monetization coefficient) increases year by year from 50 yuan / ton in the first year to an average of 75 yuan / ton in 20 years. Then, we can calculate the environmental benefit indicator cash flow:

[0140] ECFV = 80,000 tons / year * 75 yuan / ton * 20 years = 120 million yuan

[0141] This means that considering only the environmental benefit of carbon emission reduction, the project will create an environmental and economic value of approximately 120 million yuan during its 20-year lifespan.

[0142] It should be noted that this is just a simplified example. In actual applications, we need to consider more environmental benefit indicators and make more accurate predictions of annual environmental benefits and monetization coefficients. In addition, we also need to consider the changes in these values ​​under different scenarios. For example, under a stricter environmental policy scenario, the carbon price will be higher, resulting in higher environmental economic benefits.

[0143] Through this method, we can transform the environmental benefits of a project into a quantifiable and comparable economic value. This not only helps to more comprehensively evaluate the overall value of the project, but also provides an important basis for the pricing and risk assessment of green finance projects. In addition, this method also helps to identify and promote projects with significant environmental benefits, thereby promoting the healthy development of green finance.

[0144] In another preferred embodiment of the present invention, step S5 is further refined into five sub-steps, aiming to comprehensively evaluate the environmental benefits of the project and its impact on the financial feasibility of the project. The specific implementation methods of these sub-steps are as follows:

[0145] First, in step S5.1, the environmental assessment unit 33 receives environmental data from the model simulation unit 13 and determines whether the data meets the preset threshold requirements. The environmental data includes carbon emission reduction, pollutant emission reduction, ecosystem impact and other aspects. The setting of thresholds is usually based on national or local environmental standards and industry best practices.

[0146] For example, for a biomass power generation project, we will set the following thresholds: Carbon emission reduction: at least 0.8 tons of carbon dioxide per megawatt-hour of power generation; Dust emission: no more than 10mg / m 3 ; Nitrogen oxide emissions: no more than 50mg / m 3 ; Biodiversity impact: The number of species in the project area will not decrease by more than 5%;

[0147] Secondly, in step S5.2, the environmental evaluation unit 33 will provide corresponding environmental benefits according to the judgment result. If an environmental indicator meets or exceeds the threshold requirement, a positive environmental benefit will be generated; on the contrary, if the threshold is not reached, a negative environmental benefit will be generated. The environmental benefit is calculated using the following formula:

[0148] EB=∑(EI i ·W i )

[0149] Among them, EB is environmental benefit, EI is i is the value of the i-th environmental indicator, W i is the weight of the i-th environmental indicator. This formula allows us to consider multiple environmental indicators comprehensively and weight them according to their relative importance.

[0150] For example, suppose we focus on three main environmental indicators: carbon emission reduction (EI1), air quality improvement (EI2), and water conservation (EI3). We would assign weights like this: W1 (carbon emission reduction) = 0.5; W2 (air quality) = 0.3; W3 (water conservation) = 0.2;

[0151] If the project's performance on these three indicators is: EI1 = 1 million tons / year (exceeding the threshold); EI2 = 80% improvement (reaching the threshold); EI3 = 5% savings (not reaching the threshold, assuming the threshold is 10%);

[0152] Then, the calculation result of environmental benefit EB is:

[0153] EB=100*0.5+80*0.3+5*0.2=75.5

[0154] This result can be used to quantify the overall environmental benefits of the project.

[0155] Then, in steps S5.3 and S5.4, we add the positive and negative environmental benefits to the previously calculated NPV and NPV0 cash flows, respectively. The purpose of this is to monetize the environmental benefits and include them in the overall financial evaluation of the project.

[0156] Specifically, in step S5.3, we add the positive environmental benefits to NPV and NPV0 to form the positive environmental benefit cash flow NPV1 and the positive environmental benefit net present value NPV1'. For example, if the original NPV is 10 million yuan and the positive environmental benefit present value is 2 million yuan, then NPV1 will become 12 million yuan.

[0157] Similarly, in step S5.4, we add the negative environmental benefits to NPV and NPV0 to form negative environmental benefit cash flow NPV2 and negative environmental benefit net present value NPV2'. For example, if the original NPV is 10 million yuan and the negative environmental benefit present value is 1 million yuan, then NPV2 will become 9 million yuan.

[0158] Finally, in step S5.5, we compare the positive environmental benefit NPV1' and the negative environmental benefit NPV2' with the project investment cost cash flow CI. The purpose of this step is to assess whether the project is still financially viable after taking into account the environmental benefits.

[0159] The judging criteria are as follows:

[0160] If NPV1'>NPV2', the project is considered to be financially balanced, that is, the project is still feasible after considering environmental benefits.

[0161] If NPV1' < NPV2', the project is considered to be financially unbalanced, i.e. the project is not financially viable after considering environmental benefits.

[0162] The advantage of this approach is that it comprehensively considers the environmental impact of the project, including both positive and negative benefits. By monetizing environmental benefits and incorporating them into financial assessments, we can more accurately assess the true value and feasibility of green finance projects. This will not only help make more informed investment decisions, but also incentivize projects that are truly beneficial to the environment and promote the healthy development of green finance.

[0163] In a further optimized embodiment of the present invention, we add a sub-step S5.6 to step S5. The purpose of this step is to more accurately evaluate the financial feasibility of the project by calculating the capital balance rate and provide a comprehensive project environmental financial analysis report.

[0164] The calculation formula of capital balance rate (BR) is as follows:

[0165] BR=(NPV1NPV2) / CI

[0166] Among them, BR is the capital balance rate, NPV1 is the positive environmental benefit cash flow, NPV2 is the negative environmental benefit cash flow, and CI is the investment cost cash flow.

[0167] This ratio actually measures the net environmental benefits of a project (positive benefits minus negative benefits) relative to the initial investment. It can help us better understand when a project's environmental benefits are small relative to the size of its investment.

[0168] Based on this balance rate, we can provide a detailed environmental financial analysis report for the project. This report usually includes the following parts: Feasibility analysis of green finance projects: In this example, the balance rate is 120%, which is much higher than 100%, indicating that the project is not only financially feasible, but also can create significant environmental value. We can conclude that the project is highly feasible. Environmental benefit analysis plan: The report will analyze the environmental benefits of the project in detail, including positive benefits (such as carbon emission reduction, renewable energy production, etc.) and negative effects (such as potential impact on marine ecosystems). In this example, the positive benefits (1.5 billion yuan) are much greater than the negative effects (300 million yuan), indicating that the project has a positive impact on the environment as a whole. Return on investment analysis: The return on investment after taking into account environmental benefits can be calculated by the following formula: ROI = (NPV1NPV2CI) / CI*100% = (1.5 billion 300 million 1 billion) / 1 billion*100% = 20%.

[0169] This means that, after accounting for environmental benefits, the project has a 20% return on investment, which is generally considered a good rate of return.

[0170] Based on different capital balance rates, we will draw different conclusions and suggestions:

[0171] If BR > 100%: The project is not only financially viable but also creates significant environmental value and should be given priority.

[0172] If 50% < BR < 100%: The project is financially viable and can create a certain amount of environmental value and can be considered for implementation.

[0173] If 0% < BR < 50%: The project is financially viable, but the environmental benefits are relatively small and need further optimization.

[0174] If BR < 0%: The project is not financially viable, or the environmental costs are greater than the benefits, and needs to be re-evaluated or abandoned.

[0175] In our example, BR = 120%, far higher than 100%, so we can strongly recommend implementing this project in the report.

[0176] The advantage of this method is that it provides a simple and intuitive indicator to evaluate the environmental and financial performance of projects. By calculating the balance rate of funds, we can quickly compare different green finance projects, helping decision-makers select projects that not only have good financial returns but also create significant environmental value. This not only helps to achieve a balance between economic and environmental benefits but also promotes the sustainable development of green finance.

[0177] In an important improvement of the present invention, we introduce a model evaluation and correction part, namely step S6. The purpose of this step is to enable our evaluation system to adapt to the changing market environment and environmental conditions through continuous learning and adjustment, thereby maintaining the accuracy and effectiveness of the model.

[0178] The model evaluation and correction part mainly includes two aspects of updates: parameter update and update of conditional probabilities in the scenario conditions.

[0179] Among them, the parameter update uses the gradient descent method, and its formula is:

[0180]

[0181] Among them, θ new is the updated parameter, θ old is the parameter before update, α is the learning rate, is the gradient of the loss function with respect to the parameter.

[0182] Generally speaking, this model evaluation and correction mechanism makes our green finance project financial feasibility evaluation system a dynamic and adaptive system. It can continuously optimize itself over time, improve prediction accuracy, and thus provide more reliable support for the decision-making of green finance projects.

[0183] In actual applications, we perform this model evaluation and correction step regularly (such as monthly or quarterly), or trigger updates when major market changes or environmental events occur. This ensures that the model always reflects the latest market and environmental conditions, while not causing system instability due to too frequent updates.

[0184] Finally, based on the above system and method, a project financial analysis report is given. This report is the final output of the entire evaluation process and provides decision makers with a comprehensive and in-depth project analysis. The report mainly includes four parts:

[0185] 1. Project environmental benefit analysis report. This part of the report focuses on the environmental benefit indicators of the project and evaluates whether these indicators meet the preset threshold requirements. Environmental benefit indicators include but are not limited to: carbon emission reduction; renewable energy output; energy efficiency improvement; pollutant emission reduction; ecosystem impact. For example, for a large solar power generation project, we will set the following thresholds: annual carbon emission reduction ≥ 100,000 tons; power generation efficiency ≥ 18%; land use efficiency ≥ 1MW / hectare; the report will list the project's performance on each indicator in detail and compare it with the threshold. If an indicator fails to meet the threshold, the report will make corresponding improvement suggestions.

[0186] 2. Project feasibility report: This part of the report is based on the positive environmental benefit net present value NPV1' and the negative environmental benefit net present value NPV2' to determine the project's financial balance. The criteria are as follows:

[0187] If NPV1'>NPV2', the project is financially balanced;

[0188] If NPV1'<NPV2', the project is financially unbalanced;

[0189] For example, suppose the analysis results for a wind power project are as follows:

[0190] NPV1' (net present value of positive environmental benefits) = 150 million yuan;

[0191] NPV2' (negative net present value of environmental benefits) = 80 million yuan;

[0192] Since 150 million > 80 million, we can judge that the project is financially balanced, that is, the project is still financially feasible after considering the environmental benefits.

[0193] 3. Environmental benefit analysis plan. If the project is judged to be financially unbalanced, this part of the report will propose changes in investment strategies and adjustments in investment costs to change the situation of project financial imbalance. This includes: adjusting the project scale; adopting more efficient technologies; finding additional financing channels; applying for government subsidies or tax incentives; optimizing operating strategies to improve efficiency.

[0194] For example, if a biomass energy project results in financial imbalance due to excessively high raw material costs, the report will recommend expanding the raw material sources or investing in the development of more efficient energy conversion technologies.

[0195] 4. Return on investment analysis. The return on investment is calculated using the following formula in this section:

[0196] ROI = (NPV1' / CI) * 100%;

[0197] Where ROI is the return on investment, NPV1' is the net present value of positive environmental benefits, and CI is the investment cost.

[0198] This indicator reflects the overall return level of the project considering environmental benefits. For example, assume the data of a hydropower project is as follows:

[0199] NPV1' (net present value of positive environmental benefits) = 200 million yuan;

[0200] CI (investment cost) = 150 million yuan;

[0201] Then, the return on investment is:

[0202] ROI = (200 million / 150 million) * 100% ≈ 133.33%

[0203] This means that considering environmental benefits, the return rate of the project is approximately 133.33%, which is generally considered a very substantial return.

[0204] The report will evaluate the return level of the project based on the calculated ROI. Generally speaking, we will use the following evaluation criteria:

[0205] ROI > 200%: Extremely high return, a very attractive project;

[0206] 150% < ROI ≤ 200%: High return, a very attractive project;

[0207] 100% < ROI ≤ 150%: Good return, a project worthy of investment;

[0208] 50% < ROI ≤ 100%: Average return, a project that can be considered;

[0209] 0% < ROI ≤ 50%: Low return, requires careful evaluation;

[0210] ROI ≤ 0%: Negative return, usually not recommended for investment;

[0211] In our example, an ROI of 133.33% will be evaluated as a project with good return and worthy of investment.

[0212] Through such a comprehensive financial analysis report, decision makers can fully understand the project from multiple perspectives, including environmental benefits, financial feasibility, risk management and investment returns. This will not only help make more informed investment decisions, but also promote the healthy development of green finance projects and advance the realization of environmental protection and sustainable development goals.

[0213] In summary, the financial feasibility assessment system and method for green finance projects based on machine learning proposed in the present invention integrates financial assessment, market assessment and environmental assessment, and uses machine learning technology for dynamic optimization, providing a comprehensive, accurate and flexible solution for the assessment of green finance projects. This method not only takes into account traditional financial indicators, but also monetizes environmental benefits and incorporates them into the assessment system, truly realizing the unified consideration of economic and environmental benefits. Through this innovative assessment method, we can better identify and support projects that have both economic value and significant environmental benefits, thereby promoting the healthy development of green finance and contributing to addressing climate change and achieving sustainable development.

[0214] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A green finance project financial feasibility assessment system based on machine learning, characterized by: It includes a financial evaluation module, a market evaluation module and an environmental evaluation module; the financial evaluation module includes an analysis and evaluation unit, a prediction and analysis unit and a model simulation unit; the analysis and evaluation unit is used to collect and organize project and customer data and process the data; the prediction and analysis unit uses an optimization algorithm to predict financial indicators under different market scenarios based on customer data and market data; the model simulation unit provides risk scenario simulation and cash flow simulation of recovery income under different scenarios; the market evaluation module includes a data analysis unit and a prediction and analysis unit, the data analysis unit collects market data and industry data, and screens them according to the classification of green financial projects; the prediction and analysis unit predicts the changes in technical parameters of green projects under different market prices or environmental policy scenarios under future scenarios based on the screened data; the environmental evaluation module includes a data analysis unit, a model simulation unit and an environmental evaluation unit, the data analysis unit collects and analyzes environmental data of the project location; the model simulation unit performs numerical modeling on environmental data; the environmental evaluation unit judges the acceptability of environmental benefits by comparing the simulated environmental benefit indicators with the set environmental benefit indicator thresholds.

2. The green finance project financial feasibility assessment system based on machine learning according to claim 1 is characterized in that: The market assessment module and the environmental assessment module share analysis data in the following manner: DS={d|d∈M∩E} Among them, DS is the shared dataset, M is the market assessment module dataset, and E is the environmental assessment module dataset; The market assessment module uses a screening algorithm to match market data with green finance projects. The screening algorithm uses the following formula: S(pm=∑(w i ·sin(p i ,m i )) Among them, S(p,m) is the matching score between project p and market data m, w i is the weight of the i-th feature, sim(p i ,m i ) is the similarity function of the i-th feature.

3. The green finance project financial feasibility assessment system based on machine learning according to claim 1 is characterized in that: The forecasting and analysis unit and the model simulation unit jointly evaluate the impact of changes in the market environment and the risk of recovery income on cash flow. The evaluation method adopts the following formula: Among them, CF is the net present value of cash flow, R t is the profit at time t, C t is the cost at time t, and r is the discount rate. The environmental data of the market assessment module is provided by the environmental assessment module, and the data transmission is encrypted: Among them, E(D) is the encrypted data, D is the original data, K is the key, It is an XOR operation.

4. According to claim 1, the method of the financial feasibility assessment system of green finance projects based on machine learning comprises the following steps: S1, collecting market data, project data and environmental data of green finance projects, and classifying customers according to the market data and project data to achieve customer classification; S2, the prediction and analysis unit uses an optimization algorithm to predict the changes in different financial indicators of customers under market conditions; S3, the model simulation unit establishes different risk scenarios for the changes in different financial indicators of customers, and calculates the net present value (NPV) of the project through numerical simulation of the recovery income cash flow; S4, the model simulation unit performs numerical modeling on the changes in different technical indicators of the project under environmental changes, and finally obtains the technical recovery indicator cash flow; S5, the environmental evaluation unit evaluates whether different environmental benefit indicators meet the predetermined threshold.

5. The method according to claim 4, characterized in that Step S1 includes: S1.1, the market collection unit regularly collects market data and industry data; S1.2, the market collection unit classifies green finance projects; S1.3, according to the classification of green finance project types, matches the project data of customers; S1.4, collects local environmental data; S1.5, the environmental assessment module screens, organizes and cleans the collected environmental data; Step S2 includes: S2.1, the prediction and analysis unit provides several market scenario conditions, including market scenario parameters and scenario condition probabilities; scenario parameters include market electricity prices, market coal prices, market natural gas prices, and market environmental protection policies; S2.2, the prediction and analysis unit randomly generates several market scenario conditions according to the scenario parameters; S2.3, according to the generated market scenario conditions, obtains several technical benefits under market scenarios from the model simulation unit; S2.4, based on the generated market scenario conditions and the obtained several technical benefits, the prediction and analysis unit optimizes several of the technical benefits using an optimization algorithm to obtain several financial indicators under different market scenarios; S2.5, the model simulation unit receives several financial indicators under different market scenarios.

6. The method according to claim 4, characterized in that Step S3 further includes: S3.1, the model simulation unit receives financial indicators under several different market scenarios, and sets risk scenario parameters and risk scenario conditional probabilities; S3.2, the model simulation unit establishes several risk scenarios based on risk scenario parameters and risk scenario conditional probabilities for the financial indicators under several different market scenarios; S3.3, the model simulation unit calculates the project recovery income cash flow NPV0 according to several risk scenarios; S3.4, the model simulation unit simulates several risk scenarios and NPV0 to generate several recovery income cash flow NPVs, wherein the NPV calculation formula is as follows: Among them, CF t is the cash flow at time t, r is the discount rate, and I0 is the initial investment amount.

7. The method based on machine learning according to claim 4, characterized in that: Step S4 further includes: S4.1, the model simulation unit receives technical indicators under several different market scenarios; S4.2, the model simulation unit receives different technical indicators and performs modeling to obtain environmental economic benefit indicators under several different scenarios; S4.3, the model simulation unit simulates the corresponding environmental benefit indicator cash flow according to the environmental economic benefit indicators under several different scenarios, wherein the environmental benefit indicator cash flow calculation formula is as follows: ECFV=∑(EB t ·P t ) Among them, ECFV is the environmental benefit indicator cash flow, EB t is the environmental benefit at time t, P t is the environmental benefit monetization coefficient at time t.

8. The method based on machine learning according to claim 6, characterized in that: Step S5 further includes: S5.1, the environmental assessment unit receives environmental data from the model simulation unit and determines whether the environmental data meets the threshold requirement; S5.2, the environmental assessment unit provides corresponding environmental benefits according to the judgment result; according to whether the judgment result is met, positive environmental benefits and negative environmental benefits are provided respectively; S5.3, the positive environmental benefits are added to the recovery income cash flow NPV and the net present value NPV0 respectively to form a positive environmental benefit cash flow NPV1 and a positive environmental benefit net present value NPV1'; S5.4, the negative environmental benefits are added to the recovery income cash flow NPV and the net present value NPV0 respectively to form a negative environmental benefit cash flow NPV2 and a negative environmental benefit net present value NPV2'; S5.5, the positive environmental benefit net present value NPV1' and the negative environmental benefit net present value NPV2' are compared with the project investment cost cash flow CI. If NPV1'>NPV2', the project is financially balanced; if NPV1'<NPV2', the project is financially unbalanced, wherein the environmental benefit calculation formula is as follows: EB=∑(NO i ·W i ) Among them, EB is environmental benefit, EI is i is the value of the i-th environmental indicator, W i is the weight of the i-th environmental indicator.

9. The method according to claim 8, characterized in that Step S5 also includes S5.6, S5.6, comparing the positive environmental benefit cash flow NPV1 and the negative environmental benefit cash flow NPV2 with the project's investment cost cash flow CI, calculating the project's capital balance rate, and providing a project environmental financial analysis report based on different capital balance rates, including the feasibility of the green finance project, the environmental benefit analysis plan and the rate of return on investment, wherein the capital balance rate calculation formula is as follows: BR=(NPV1-NPV2) / CI Among them, BR is the capital balance rate, NPV1 is the positive environmental benefit cash flow, NPV2 is the negative environmental benefit cash flow, and CI is the investment cost cash flow.

10. The method according to claim 9, characterized in that It also includes S6, model evaluation and correction part; the model evaluation and correction part corrects the model simulation unit in real time according to market and environmental data; the adjustment of the model simulation unit includes the update of parameters and the update of conditional probabilities in scenario conditions, wherein the parameter update adopts the following formula: Among them, θ new is the updated parameter, θ old is the parameter before updating, α is the learning rate, is the gradient of the loss function with respect to the parameters.

Images