Water footprint evaluation method and equipment

Abstract

The embodiment of the invention provides a water footprint evaluation method and equipment. In the embodiment of the invention, on one hand, multiple evaluation dimensions such as attributes of raw material crops, a geographical environment and a production environment of aviation fuel are introduced to serve as evaluation dimensions for performing water footprint evaluation on the aviation fuel; and on the other hand, a random forest regression model is adopted, and water footprint evaluation is carried out based on the introduced characteristic parameters under each evaluation dimension. Therefore, by introducing the random forest regression model and the multi-dimensional characteristic parameters, the complex influence of various characteristic parameters on water footprint evaluation can be fully analyzed, so that the accuracy of water footprint evaluation is improved; moreover, the method can adapt to the diversity of raw material crops, geographical environments and production environments, and the water footprint of the sustainable aviation fuel under different characteristic parameter combinations can be efficiently evaluated by flexibly adjusting the input characteristic parameter set, so that rich reference is provided for the production of the sustainable aviation fuel; and furthermore, the water footprint of the sustainable aviation fuel is better optimized, and the goal of sustainable development is accelerated.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a water footprint assessment method and device. Background Art

[0002] With the rapid development of the aviation transportation industry, the demand for aviation fuel continues to grow. However, the use of traditional fossil fuels has led to a large amount of carbon emissions and environmental pollution. Therefore, the development of low-carbon and environmentally friendly sustainable aviation fuel (SAF) has become one of the important directions to solve this problem. SAF should not only reduce carbon emissions, but also reduce other environmental impacts. Water footprint is one of the important indicators to measure the impact of fuel production on the environment.

[0003] At present, the water footprint of sustainable aviation fuel is usually assessed by "direct measurement method" or "water consumption rate method". The direct measurement method is to directly measure the amount of water resources used in each water-using link involved in sustainable aviation fuel and calculate the water footprint; while the water consumption rate method is to predict the required water footprint based on historical data or industry averages, combined with production scale. However, the direct measurement method often has problems such as missed measurements and wrong measurements, and the basic parameter values ​​used in the water consumption rate method are too one-sided, resulting in poor accuracy of the water footprint assessed by the existing assessment methods. Summary of the invention

[0004] Multiple aspects of the present application provide a water footprint assessment method and device to improve the accuracy of water footprint assessment for sustainable aviation fuel.

[0005] The present application provides a water footprint assessment method, which includes:

[0006] In response to a water footprint assessment instruction initiated for sustainable aviation fuel to be produced, a feature parameter set to be processed is obtained, wherein the feature parameter set includes attribute feature parameters of raw material crops, geographical environment feature parameters of raw material crops, and / or production environment feature parameters of sustainable aviation fuel;

[0007] Inputting the characteristic parameter set into a pre-trained water footprint assessment model, wherein the water footprint assessment model adopts a random forest regression model, wherein the water footprint assessment model includes a plurality of preset decision trees, and the water footprint assessment model is used to perform water footprint assessment on the characteristic parameter set respectively through the plurality of decision trees;

[0008] In the water footprint assessment model, the characteristic parameter sets are respectively input into the multiple decision trees, so as to generate assessment results respectively through the multiple decision trees;

[0009] The evaluation results generated by the multiple decision trees are aggregated to obtain a final evaluation result as the water footprint evaluation result corresponding to the sustainable aviation fuel.

[0010] Furthermore, the characteristic parameter set is input into the pre-trained water footprint assessment model, including:

[0011] Dividing the characteristic parameters contained in the characteristic parameter set into at least one parameter group, where different parameter groups correspond to different variable types;

[0012] If there is a first parameter group whose variable type is a categorical variable, one-hot encoding is performed on the feature parameters in the first parameter group to obtain corresponding preprocessed parameters;

[0013] If there is a second parameter group whose variable type is a numerical variable, standardizing the characteristic parameters in the second parameter group to obtain corresponding preprocessed parameters;

[0014] The preprocessed characteristic parameter set is input into the water footprint assessment model.

[0015] Furthermore, the attribute parameters of the raw material crops include parameter values ​​under one or more characteristic dimensions of crop category, growth cycle, unit yield and reference water footprint; the geographical environment parameters of the raw material crops include parameter values ​​under one or more characteristic dimensions of geographical location, precipitation and sunshine duration; the production environment parameters of the sustainable aviation fuel include parameter values ​​under one or more characteristic dimensions of raw material transportation distance and energy proportion; wherein, the first parameter group includes parameter values ​​under the crop category and / or the geographical location; and the second parameter group includes parameter values ​​under the growth cycle, the unit yield, the reference water footprint, the precipitation, the sunshine duration, the transportation distance and / or the energy proportion.

[0016] Furthermore, the method further comprises:

[0017] In the process of performing water footprint assessment on the characteristic parameter set by the multiple decision trees in the water footprint assessment model respectively, determining the impact of the assessment result corresponding to each node included in the multiple decision trees, each node corresponding to a characteristic dimension;

[0018] Based on the impact of the evaluation results determined for each node, the feature importance is statistically analyzed in units of feature dimensions;

[0019] Among them, in the multiple decision trees, there are one or more nodes corresponding to the same feature dimension. If there are multiple nodes, the influence of the evaluation results corresponding to the multiple nodes is fused to calculate the feature importance corresponding to the feature dimension.

[0020] Furthermore, the method further comprises:

[0021] Sort the feature dimensions according to the feature importance calculated for different feature dimensions;

[0022] The feature importance of the sorted feature dimensions is visualized to show the contribution of different feature dimensions to the water footprint assessment results.

[0023] Furthermore, the training process of the water footprint assessment model includes:

[0024] Under the preset evaluation dimensions, characteristic dimensions are respectively determined to obtain a characteristic dimension set, wherein the evaluation dimensions include one or more factors in the attributes of the raw material crop, the geographical environment of the raw material crop, and / or the production environment of the sustainable aviation fuel;

[0025] For any decision tree to be constructed, select some feature dimensions from the feature dimension set;

[0026] Extract multiple training samples from the training sample set as a training set;

[0027] Based on the selected feature dimensions, construct the decision tree, wherein the feature dimensions serve as nodes in the decision tree;

[0028] Continue to construct other decision trees to form the water footprint assessment model based on the constructed multiple decision trees;

[0029] Based on the training set, the model parameters of the water footprint assessment model are optimized and the decision tree structure is adjusted to complete the training of the water footprint assessment model.

[0030] Furthermore, the method further comprises:

[0031] Extracting multiple training samples from the training sample set as a test set;

[0032] Inputting the test set into the water footprint assessment model to obtain a test result;

[0033] Performing performance verification on the water footprint assessment model in a cross-validation manner to optimize the hyperparameters corresponding to the water footprint assessment model;

[0034] The validation indicators used in the cross-validation process include mean square error and / or mean absolute error.

[0035] Furthermore, the evaluation results generated by the multiple decision trees are aggregated to obtain a final evaluation result, including:

[0036] Calculating the mean of the evaluations generated by the plurality of decision trees;

[0037] The calculated mean is taken as the final evaluation result.

[0038] Furthermore, the method further comprises:

[0039] The water footprint assessment results corresponding to the sustainable aviation fuel are visualized.

[0040] The embodiment of the present application also provides a computing device, including a memory, a processor, and a communication component;

[0041] The memory is used to store one or more computer instructions;

[0042] The processor is coupled to the memory and the communication component, and is configured to execute the one or more computer instructions to perform the aforementioned water footprint assessment method.

[0043] In the embodiment of the present application, a water footprint assessment method for sustainable aviation fuel is proposed. On the one hand, it is proposed to introduce multiple assessment dimensions such as the properties of raw material crops, geographical environment, and the production environment of sustainable aviation fuel as assessment dimensions for water footprint assessment of sustainable aviation fuel; on the other hand, it is also proposed to use a random forest regression model to conduct water footprint assessment based on the characteristic parameters under each assessment dimension introduced. In this way, by introducing a random forest regression model and multi-dimensional characteristic parameters, the complex impact of various characteristic parameters on water footprint assessment can be fully analyzed, thereby improving the accuracy of water footprint assessment; moreover, it can also adapt to the diversity of raw material crops, geographical environment, and production environment, and by flexibly adjusting the input characteristic parameter set, the water footprint of sustainable aviation fuel under different characteristic parameter combinations can be efficiently assessed, thereby providing a rich reference for the production of sustainable aviation fuel, thereby better optimizing the water footprint of sustainable aviation fuel and accelerating the realization of sustainable development goals. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0045] Figure 1 A schematic flow chart of a water footprint assessment method provided for this application;

[0046] Figure 2 A schematic diagram of a process flow related to data collection and preprocessing provided for this application;

[0047] Figure 3A schematic diagram of a model training and verification process provided in this application;

[0048] Figure 4 A schematic diagram of a feature importance visualization effect provided for this application;

[0049] Figure 5 A schematic diagram of the structure of a computing device is provided for yet another exemplary embodiment of the present application. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0051] Before starting to explain in detail the technical solutions provided by the embodiments of the present application, several technical concepts involved in the present application are briefly explained as follows.

[0052] Water footprint can be understood as the amount of water resources required for all products and services consumed by a country, a region or an individual within a certain period of time.

[0053] Sustainable aviation fuel is a special fuel designed specifically for aircraft.

[0054] As introduced in the background technology, the water footprint of sustainable aviation fuel is usually assessed by “direct measurement method” or “water consumption rate method”. However, the water footprint assessed by these existing assessment methods is not accurate.

[0055] In order to overcome the above problems existing in the prior art, this embodiment proposes to combine the fields of environmental science, aviation new energy and artificial intelligence to realize a water footprint assessment method for sustainable aviation fuel, so as to improve the accuracy of water footprint assessment for sustainable aviation fuel. The technical concept of the embodiment of this application is described below.

[0056] The technical concept of the embodiments of the present application proposes: introducing multiple assessment dimensions such as the properties of raw material crops, the geographical environment, and the production environment of sustainable aviation fuel as assessment dimensions for water footprint assessment of sustainable aviation fuel.

[0057] After research and exploration, the inventors found that the various raw material crops used to produce sustainable aviation fuel, such as reed and castor, have a significant impact on the environment in terms of water consumption (i.e., water footprint) during the growth and production of sustainable aviation fuel. It is proposed to design a rich feature dimension under the two evaluation dimensions of the properties of the raw material crops and the geographical environment to explore the complex impact of these two evaluation dimensions on the water footprint. The inventors also found that the production environment of sustainable aviation fuel also has an important impact on the water footprint. Therefore, it is also proposed to design a rich feature dimension under the evaluation dimension of the production environment of sustainable aviation fuel to explore the complex impact of this evaluation dimension on the water footprint.

[0058] For example, under the evaluation dimension of the attributes of the raw material crops in the embodiment of the present application, the introduced characteristic dimensions may include but are not limited to crop category, growth cycle, unit yield and reference water footprint, etc. Under the evaluation dimension of the address environment of the raw material crops, the introduced characteristic dimensions may include but are not limited to geographical location, precipitation and sunshine duration, etc. Under the evaluation dimension of the production environment of sustainable aviation fuel, the introduced characteristic dimensions may include but are not limited to transportation distance and energy proportion, etc., where the energy proportion is used to describe the proportion of different types of energy in the production process of sustainable aviation fuel, for example, the energy proportion occupied by coal power, the proportion occupied by natural gas power generation and the proportion occupied by renewable energy. Raw material crops are typical renewable resources.

[0059] The technical concept of the embodiment of the present application also proposes: using a random forest regression model to perform water footprint assessment based on the characteristic parameters under each evaluation dimension introduced.

[0060] Among them, the Random Forest Regression model is an ensemble learning algorithm based on decision trees, which is used to solve regression problems. It constructs multiple decision trees and integrates the prediction results of these decision trees to obtain the final regression prediction value. The basic idea of ​​ensemble learning is to combine multiple weak learners (the weak learners here are decision trees) to form a strong learner to improve the prediction performance and stability of the model.

[0061] Based on the introduction of the random forest regression model and the above-mentioned rich feature dimensions, the water footprint assessment method provided in the embodiments of the present application can fully take into account the various complex factors affecting the water footprint, such as geographical location, precipitation, sunshine hours, transportation distance and crop characteristics, and fully explore the nonlinear relationship between these complex factors and the water footprint, thereby fully analyzing the complex impact of these complex factors on the water footprint, thereby effectively improving the accuracy of water footprint assessment.

[0062] The technical concept of the embodiments of the present application is described in detail below with reference to the accompanying drawings.

[0063] Figure 1 This is a flow chart of a water footprint assessment method provided by the present application. The method can be executed by a data processing device. The data processing device can be implemented as software, hardware, or a combination of software and hardware. The data processing device can be integrated in a computing device. Figure 1 , the method may include:

[0064] Step 100: In response to a water footprint assessment instruction initiated for sustainable aviation fuel to be produced, a feature parameter set to be processed is obtained, where the feature parameter set includes attribute feature parameters of raw material crops, geographical environment feature parameters of raw material crops, and / or production environment feature parameters of sustainable aviation fuel;

[0065] Step 101: input the feature parameter set into a pre-trained water footprint assessment model, where the water footprint assessment model adopts a random forest regression model. The water footprint assessment model includes a plurality of preset decision trees, and the water footprint assessment model is used to perform water footprint assessment on the feature parameter set through the plurality of decision trees;

[0066] Step 102: In the water footprint assessment model, the characteristic parameter sets are respectively input into a plurality of decision trees, so as to generate assessment results through the plurality of decision trees respectively;

[0067] Step 103: Aggregate the evaluation results generated by the multiple decision trees to obtain a final evaluation result as the water footprint evaluation result corresponding to the sustainable aviation fuel.

[0068] In this embodiment, there is no limitation on the timing of initiating the water footprint assessment instruction in step 100. When water footprint assessment needs to be performed based on different characteristic parameter sets, the water footprint assessment instruction can be initiated as needed.

[0069] The characteristic parameter set in this embodiment includes several parameter values. This corresponds to the multiple evaluation dimensions introduced above and the characteristic dimensions designed under each evaluation dimension. The characteristic parameter set includes the parameter values ​​under each characteristic dimension introduced in this embodiment.

[0070] Following the feature dimensions in the previous example, in step 100, in the feature parameter set, the attribute parameters of the raw material crops include parameter values ​​under one or more feature dimensions of crop category, growth cycle, unit yield and reference water footprint; the geographical environment parameters of the raw material crops include parameter values ​​under one or more feature dimensions of geographical location, precipitation and sunshine duration; the production environment parameters of sustainable aviation fuel include parameter values ​​under one or more feature dimensions of transportation distance and energy proportion. It can be understood that these feature dimensions are only exemplary, and this embodiment is not limited thereto. More feature dimensions can also be designed based on the evaluation dimensions introduced in this embodiment to more comprehensively mine the feature dimensions that affect the water footprint. More feature dimensions are not given here.

[0071] Continue to refer Figure 1 In step 101, the feature parameter set may be input into the pre-trained water footprint assessment model. The water footprint assessment model in this embodiment adopts a random forest regression model. In this embodiment, the input of the water footprint assessment model is the feature parameter set, and the output is the assessment result.

[0072] refer to Figure 1 After the characteristic parameter set is input into the water footprint assessment model, in step 102, the characteristic parameter set will enter the multiple decision trees contained in the water footprint assessment model respectively. The multiple decision trees can independently perform water footprint assessment based on the characteristic parameter set and generate assessment results respectively.

[0073] In step 103, the evaluation results generated by the multiple decision trees may be aggregated to obtain a final evaluation result as a water footprint evaluation result corresponding to the sustainable aviation fuel.

[0074] In step 103, an optional aggregation operation may be: averaging the evaluations generated by multiple decision trees; and using the calculated average as the final evaluation result. Of course, the aggregation operation may also be implemented by taking the median, etc., which is not limited here, and no further examples are given.

[0075] It is worth noting that the aggregation operation in step 103 can be completed in the water footprint assessment model. Of course, the data processing device in this embodiment can also obtain the assessment results generated by multiple decision trees from the water footprint assessment model and then perform the aggregation operation. This embodiment does not limit this.

[0076] In addition, the technical concept of the embodiment of the present application also proposes that the water footprint assessment results corresponding to sustainable aviation fuel can be visualized.

[0077] In summary, in the technical concept of this embodiment, a water footprint assessment method for sustainable aviation fuel is proposed. On the one hand, it is proposed to introduce multiple assessment dimensions such as the properties of raw material crops, geographical environment, and the production environment of sustainable aviation fuel as the assessment dimensions for water footprint assessment of sustainable aviation fuel; on the other hand, it is also proposed to use a random forest regression model to conduct water footprint assessment based on the characteristic parameters under each assessment dimension introduced. In this way, by introducing a random forest regression model and multi-dimensional characteristic parameters, the complex impact of various characteristic parameters on water footprint assessment can be fully analyzed, thereby improving the accuracy of water footprint assessment; moreover, it can also adapt to the diversity of raw material crops, geographical environment, and production environment. By flexibly adjusting the input characteristic parameter set, the water footprint of sustainable aviation fuel under different characteristic parameter combinations can be efficiently assessed, thereby providing a rich reference for the production of sustainable aviation fuel, thereby better optimizing the water footprint of sustainable aviation fuel and accelerating the realization of sustainable development goals.

[0078] In the technical concept of the embodiment of the present application, it is also proposed that: the characteristic parameter set can be preprocessed before being input into the water footprint assessment model. An optional technical concept proposes that: the characteristic parameters contained in the characteristic parameter set can be divided into at least one parameter group, and different parameter groups correspond to different variable types; if there is a first parameter group whose variable type is a categorical variable, the characteristic parameters in the first parameter group are uniquely encoded to obtain the corresponding preprocessed parameters; if there is a second parameter group whose variable type is a numerical variable, the characteristic parameters in the second parameter group are standardized to obtain the corresponding preprocessed parameters; the preprocessed characteristic parameter set is input into the water footprint assessment model.

[0079] In this optional technical concept, the characteristic parameters in the characteristic parameter set are grouped according to the variable type, and different preprocessing methods are designed for the different parameter groups divided. Among them, the unique hot encoding technology will be demonstrated in detail later. Standardization can be understood as unifying the dimensions of the values ​​in different parameter groups to more balancedly analyze the impact of the relevant characteristic dimensions on the water footprint assessment.

[0080] Following the exemplary characteristic dimensions above, the first parameter group and the second parameter group are exemplified here. The first parameter group may include parameter values ​​under crop category and / or geographical location; the second parameter group may include parameter values ​​under growth cycle, unit yield, reference water footprint, precipitation, sunshine duration, transportation distance and / or energy proportion. There is no limitation on the number of parameter groups grouped, the types of variables involved in different parameter groups, and the characteristic dimensions included in different parameter groups.

[0081] By preprocessing the characteristic parameter set, the characteristic parameter set can be organized in a more standardized and structured manner, thereby making it easier for the water footprint assessment model to perform data analysis.

[0082] The technical concept of the embodiment of the present application also proposes: in the process of multiple decision trees in the water footprint assessment model performing water footprint assessment on the feature parameter set respectively, the impact of the assessment results corresponding to each node contained in the multiple decision trees can be determined, and each node corresponds to a feature dimension; based on the impact of the assessment results determined for each node, the feature importance is statistically analyzed in units of feature dimensions; wherein, in the multiple decision trees, there are one or more nodes corresponding to the same feature dimension. If there are multiple nodes, the impact of the assessment results corresponding to the multiple nodes is fused to calculate the feature importance corresponding to the feature dimension.

[0083] Among them, feature importance can be used to reflect the contribution of feature dimensions to the evaluation results. Among them, the impact of the evaluation results determined for the nodes mentioned above can be expressed as Mean Decrease in Impurity or Mean Decrease in Impurity with Permutation. Taking Gini impurity as an example, for a feature dimension, at each node of each decision tree, if the feature dimension is selected for splitting, the difference between the Gini impurity before and after the split is calculated (representing "impurity reduction"), and these differences are averaged in all nodes and all decision trees to obtain the average impurity reduction value of the feature dimension. The larger this value is, the more important the feature dimension is, that is, the higher the feature importance is.

[0084] Furthermore, the feature dimensions may be sorted according to the feature importance counted for different feature dimensions; the feature importance of the sorted feature dimensions may be visualized to present the contribution of different feature dimensions to the water footprint assessment result.

[0085] By analyzing the importance of characteristics, we can fully explore the key factors affecting water footprint and provide a basis for optimizing water resources management.

[0086] The technical concept of the embodiment of the present application also proposes: pre-training the water footprint assessment model. To this end, the characteristic dimensions can be determined separately under the preset assessment dimensions to obtain a characteristic dimension set; referring to the above, the assessment dimensions may include one or more dimensions of the attributes of the raw material crops, the geographical environment of the raw material crops and / or the production environment of sustainable aviation fuel; for any decision tree to be constructed, some characteristic dimensions are selected from the characteristic dimension set; multiple training samples are extracted from the training sample set as a training set; based on the selected characteristic dimensions, a decision tree is constructed, and the characteristic dimensions are used as nodes in the decision tree; other decision trees are continued to be constructed to form a water footprint assessment model based on the constructed multiple decision trees; based on the training set, the model parameters of the water footprint assessment model are optimized and the decision tree structure is adjusted to complete the training of the water footprint assessment model.

[0087] In addition, multiple training samples can be extracted from the training sample set as a test set; the test set is input into the water footprint assessment model to obtain the test results; the water footprint assessment model is cross-validated to perform performance verification to optimize the hyperparameters corresponding to the water footprint assessment model; wherein the verification indicators used in the cross-validation process include mean square error and / or mean absolute error. The hyperparameters here may include but are not limited to the number of trees n_estimators, the maximum depth of the tree max_depth, the minimum number of samples required for leaf nodes min_samples_leaf`, etc., which are not limited here and no further examples are given.

[0088] In this way, the model parameters, decision tree structure, and hyperparameters in the water footprint assessment model can be optimized based on the training set and the test set, thereby obtaining a water footprint assessment model suitable for the embodiment of the present application.

[0089] Based on the above description of the technical concept in the embodiment of the present application, the following exemplary embodiments of the present application are provided. The steps included in the exemplary embodiments are described as follows:

[0090] 1) Data collection and preprocessing

[0091] Collect the factors that affect the water footprint, such as crop types, crop growth cycles, unit yields of crops, geographical locations considered, local average annual precipitation, local average annual sunshine hours, transportation distance from raw materials to fuel processing plants, and power generation energy structure in a specific year, etc. These influencing factors correspond to the characteristic dimensions introduced in the embodiments of this application.

[0092] Special processing is required for the two data of crop type and geographical location. The variables are preprocessed by one-hot encoding and converted into a numerical format that can be used by the model. The part about one-hot encoding will be described in detail below. The rest of the data is standardized to ensure that the model can effectively process different types of data. The model here corresponds to the water footprint assessment model introduced in the embodiment of the present application.

[0093] This embodiment sorts out the process flow of this process, and the specific process is as follows: Figure 2 As shown, Figure 2 A schematic diagram of a process flow related to data collection and preprocessing provided for this application.

[0094] 2) Model construction

[0095] This embodiment adopts the random forest regression algorithm. Random forest is an integrated learning algorithm that captures complex nonlinear relationships by training multiple decision trees. Each decision tree independently learns part of the data set, and finally improves the prediction accuracy by weighted averaging the outputs of multiple trees. Define the model input features, including crop conditions, climate conditions, transportation distance, etc. The specific feature dimensions can refer to the previous description. In order to improve the prediction performance of the model, grid search (GridSearch) and cross-validation techniques are used to optimize the hyperparameters of the random forest model to build the required model. By selecting the best hyperparameter combination, it is ensured that the model has a high generalization ability under diversified input conditions.

[0096] The specific parts of the random forest regression model are as follows:

[0097] Step 1: Randomly select a sample subset as the training set for the decision tree.

[0098] Step 2: Randomly select a part of the features (the square root of the total number of features) as the feature set of the decision tree.

[0099] Step 3: Build a decision tree based on the training set and feature set until the predetermined number of leaf nodes is reached or it cannot be split.

[0100] Step 4: Repeat the above steps to build multiple decision trees.

[0101] Step 5: For a new sample, input it into each decision tree to obtain multiple prediction results.

[0102] Step 6: Average multiple prediction results to get the final prediction result.

[0103] The algorithm formula is based on the decision tree regression model. The prediction function of each decision tree can be expressed as shown in formula (1):

[0104]

[0105] Where: K represents the Kth decision tree, x represents the input sample, J k represents the number of leaf nodes of the kth decision tree, c kj represents the predicted value of the jth leaf node of the kth decision tree, R kj Represents the sample set of leaf nodes of the jth decision tree.

[0106] The prediction function of multiple decision trees can be expressed as:

[0107]

[0108] Where: P represents the number of decision trees.

[0109] In model evaluation, the evaluation indicators that can be used include mean square error (MSE) and determination coefficient R-squared (R 2， It can also be expressed as R^2). Generally speaking, the smaller the MSE value, the better the model fits the data. 2 The closer the value is to 1, the better the model fits the data, and vice versa. The calculation formula is as follows:

[0110]

[0111] In the formula, n represents the number of samples, y i represents the true value of the i-th sample, Represents the predicted value of the i-th sample.

[0112]

[0113] Where: Represents the average of all true sample values.

[0114] 3) Model training and verification

[0115] The model is trained with the collected data, and the data is divided into a training set and a test set. The cross-validation method is used to evaluate the performance of the model. The accuracy of the model is evaluated by indicators such as mean square error (MSE). The specific training and validation process is as follows: Figure 3 As shown, Figure 3 A schematic diagram of the model training and verification process provided in this application.

[0116] 4) Feature Importance Analysis

[0117] The feature importance analysis function of the random forest model can evaluate the contribution of different input variables to the water footprint prediction. Through the analysis of feature importance, it is easy for the relevant industries engaged in SAF to identify the key factors affecting the water footprint and make reasonable control in the later production process to ensure the best economic benefits.

[0118] 5) Water footprint assessment

[0119] The trained model can predict the water footprint of SAF under specific conditions based on different input variables. The input variables of the model may include: crop growth cycle, unit yield, average annual precipitation, average annual sunshine hours, transportation distance, and the proportion of coal-fired power, natural gas and renewable energy in the energy structure. These input variables correspond to the characteristic parameter set mentioned above.

[0120] 6) Results visualization

[0121] Visualization tools can be used to display actual and predicted water footprint results, and feature importance can be presented in the form of graphs. This not only helps to understand the output of the model, but also can intuitively show how well the model supports decision-making. Figure 4 A schematic diagram of a feature importance visualization effect provided in this application.

[0122] Based on the water footprint assessment method provided in the embodiments of the present application, in the above exemplary embodiments, the following technical effects can be achieved.

[0123] 1) Processing of multi-dimensional data: This embodiment introduces the random forest algorithm, which can simultaneously handle the complex impact of different crops, climate conditions, geographical locations and energy structures on water footprint, significantly improving the prediction accuracy of the model. Random forest has good nonlinear modeling capabilities and can handle complex interactions between high-dimensional variables.

[0124] 2) Dynamic optimization: The model can adapt to changes in different geographical regions, climate conditions and energy structures through continuous updating and optimization. This dynamic optimization capability makes the method have the potential for flexible application in different regions and production conditions.

[0125] 3) Feature importance analysis: This embodiment can not only predict the water footprint, but also identify the main factors affecting the water footprint through feature importance analysis, such as the water demand of specific crops, the impact of transportation distance, the contribution of energy structure to production water use, etc. This function provides a valuable reference for SAF manufacturers in designing and optimizing production processes.

[0126] 4) Applicable to different production scenarios: This method is applicable to all kinds of sustainable aviation fuel production scenarios, including the use of a variety of different biomass raw materials and energy structures. Enterprises can flexibly adjust input variables according to different crops and production conditions to obtain accurate water footprint predictions.

[0127] Moreover, the water footprint assessment method provided in the embodiment of the present application is applicable to at least the following usage scenarios:

[0128] 1) Biomass raw material planting planning: Enterprises can select biomass raw materials with higher water resource utilization efficiency and optimize planting planning based on the water footprint of different crops. At the same time, they can also choose the appropriate planting area under the premise of planting a specific crop to ensure maximum benefits.

[0129] 2) Transportation management: By evaluating the impact of different transportation distances on water footprint, companies can rationally plan the transportation routes of raw materials and reduce unnecessary resource consumption during transportation. At the same time, the geographical location of the raw material origin and the production area can be adjusted according to actual conditions to ensure the convenience of the entire production process.

[0130] 3) Production process design: Under the background of energy structure transformation, the utilization rate of clean energy and renewable energy is gradually increasing. By understanding the impact of energy structure on water footprint, relevant companies can optimize the use of energy in production processes and reduce water consumption.

[0131] In this way, through the water footprint assessment method provided in the embodiment of the present application, enterprises can not only optimize the use of resources in the production process, but also better cope with possible future environmental regulations and standards, and promote the popularization and promotion of sustainable aviation fuels.

[0132] Hereinafter, a specific use scheme of the water footprint assessment method provided in an embodiment of the present application in an exemplary application scenario is described.

[0133] 1. Data Collection

[0134] 1) First, basic data related to SAF production can be collected. The data sources may include agricultural statistics, meteorological data, energy structure data, and transportation information. The following is a collection description of exemplary data items (corresponding to the characteristic parameters described above):

[0135] 2) Crop data: Collect data on different biomass raw materials used for SAF production. The data items include crop type (such as reed, castor, rapeseed, etc.), crop growth cycle (unit: day), and unit yield (unit: ton / hectare).

[0136] 3) Climate data: Collect meteorological data of the production site, with a focus on average annual precipitation (unit: mm) and average annual sunshine hours (unit: hours). These data can be obtained from the National Statistical Yearbook or local meteorological departments.

[0137] 4) Geographic location and transportation data: Collect the transportation distance (in kilometers) between the raw material production site and the fuel processing plant, and consider the impact of different geographical locations on water resource consumption. The geographical location is based on China's major SAF production areas, such as Xinjiang Uygur Autonomous Region, Hunan Province and Hebei Province.

[0138] 5) Energy structure data: Collect the energy types and structure proportions of the current energy industry power generation, mainly including the proportion of coal-fired power generation, the proportion of natural gas power generation and the proportion of renewable energy. These data can be obtained through the energy statistical yearbook or reports provided by power companies.

[0139] 6) Reference water footprint data: Obtain the water footprint of different biomass raw materials during planting, transportation and conversion. Water footprint data can be measured through field experiments or refer to statistical data in relevant literature.

[0140] 2. Data preprocessing

[0141] The collected data needs to be cleaned and preprocessed to ensure that it can be processed by the machine learning model. In the specific implementation process, the main data preprocessing steps include:

[0142] 1) One-hot encoding of categorical variables: For categorical variables such as crop type and geographic location, one-hot encoding is used. For example, crop type (Arundo donax, Castor, Camelina, etc.) and geographic location (Xinjiang, Hunan, Hebei, etc.) are converted into multiple binary features (0 or 1) so that the model can process categorical data.

[0143] 2) Standardization of numerical variables: Standardization is used for numerical features such as crop growth cycle, unit yield, average annual precipitation, average annual sunshine hours, transportation distance, coal-fired power ratio, natural gas ratio, and renewable energy ratio. By converting each feature value into a standard normal distribution with a mean of 0 and a variance of 1, it is ensured that features of different dimensions have similar weights during model training.

[0144] This section will explain the one-hot encoding mentioned above. One-hot encoding is a method of converting categorical variables (such as color, crop type, geographic location, etc.) into numerical form. The basic idea is that each category is represented by a set of binary numbers (0 and 1), where one category is 1 and the rest are 0.

[0145] For example, suppose there are three crop categories: Arundo donax, Castor, and Camelina. Using one-hot encoding, they will be converted into three columns, representing the three crops respectively, as shown in Table 1 - One-hot encoding data:

[0146] Crop Type Arundodis Castor Camelina Arundodis 1 0 0 Castor 0 1 0 Camelina 0 0 1

[0147] In this way, each category becomes a unique set of binary digital representations, for example, the corresponding representation for Phragmites australis is 100, and the corresponding representation for Castor oil plant is 010.

[0148] In this embodiment, the purpose of one-hot encoding is to convert non-numerical classification information such as crop types (such as reed, castor, camelina, etc.) and geographical locations (such as Xinjiang, Hunan, Hebei, etc.) into numerical format to facilitate model processing.

[0149] Without one-hot encoding, the machine learning model will not be able to process these non-numeric data. If we directly use the numbers 1, 2, and 3 to represent three crops, the model may misunderstand that there is a size relationship between these numbers (such as 1 is smaller than 2), which will affect the training results of the model. After using one-hot encoding, the model will not have this misunderstanding, and each category is independent and has no order. This can avoid misunderstandings about the priority between categories and make the model more accurate.

[0150] 3. Model construction

[0151] After data preprocessing is completed, the model building phase begins. This embodiment uses a random forest regression algorithm to predict water footprint. The specific implementation steps are as follows:

[0152] 1) Model initialization: Use the `scikit-learn` library in the Python programming language to build a random forest regression model. Random forest integrates multiple decision trees and averages the prediction results of each tree to obtain more robust prediction results.

[0153] 1.from sklearn.ensemble import RandomForestRegressor

[0154] 2.rf_model=RandomForestRegressor(random_state=42)

[0155] The first line of code above imports the RandomForestRegressor class from the ensemble module of the sklearn library. sklearn is a commonly used machine learning library in Python that provides a wealth of machine learning algorithms and tools. The ensemble module contains a variety of ensemble learning algorithms, and RandomForestRegressor is an implementation class of the random forest regression algorithm, which is used to solve regression problems and can predict numerical target variables.

[0156] The second line of code above creates an instance object rf_model of the RandomForestRegressor class and passes in the parameter random_state=42. The function of random_state is to set the random number seed so that the training and prediction results of the model are repeatable. When a fixed random number seed is set, each time the code is run, the random forest model will perform random sampling of data, feature selection and other operations during the training process based on the same random sequence, thereby ensuring that the same model training results and prediction results are obtained under the same data set and model parameters. In this way, when performing model tuning and comparing the effects of different parameter settings, it is possible to more accurately evaluate whether the performance changes of the model are caused by parameter adjustments rather than random factors.

[0157] 2) Hyperparameter optimization: In order to improve the prediction performance of the model, the grid search method is used to optimize the hyperparameters. The main optimized parameters include the number of decision trees, the maximum depth, the minimum number of samples per node, etc. The exemplary steps of hyperparameter optimization are as follows:

[0158]

[0159] Among them, 'n_estimators' represents the number of decision trees in the forest. Random forests are composed of multiple decision trees, and each tree will be different during training. The more decision trees there are, the more stable and accurate the model will usually be, but the training time will also increase. 100, 200, and 300 all represent the number of components of the tree. 'max_depth' represents the maximum depth of the decision tree, that is, the number of nodes in the longest path from the root node to the leaf node. The deeper the tree, the better the model can fit the data (especially complex data). However, if the tree is too deep, it may lead to overfitting (that is, the model performs well on the training data, but performs poorly on new data). None: means that there is no limit on the depth of the tree. The tree will continue to split until all leaf nodes are pure (or there is no possibility of further splitting). 10, 20, and 30 all represent the maximum depth limit of the tree. 'min_samples_split' represents the minimum number of samples required for a node to split. When the number of samples in a node is greater than or equal to this value, the node is allowed to split. 2: Each node requires at least 2 samples before it can continue to split. This is the default value, which usually makes the tree more complex (deeper). 5: Each node needs at least 5 samples to split, which will limit the complexity of the tree. 10: Each node needs at least 10 samples to split, so that the tree will be simpler and more regular, reducing the risk of overfitting. 'min_samples_leaf' represents the minimum number of samples required in the leaf node (the end node of the tree). This parameter controls how many samples the leaf nodes of the tree must contain at least to prevent excessive splitting and causing the tree to be too complex.

[0160] 3) Model training: A random forest model with optimized hyperparameters is used to fit the training data. The model captures the complex factors that affect water footprint by continuously learning and adjusting the decision tree structure.

[0161] 4) Feature importance analysis: The random forest algorithm has a built-in feature importance assessment function, which evaluates the impact of each feature on the prediction result and identifies which factors contribute the most to the water footprint under different crops, climate conditions, geographical locations and energy structures. In this way, companies can better understand which factors need to be optimized first.

[0162]

[0163]

[0164] Among them, 'importances': get the feature importance score of the model through 'best_rf_model.feature_importances'. Feature importance indicates the contribution of a feature to the prediction result. The higher the value, the greater the importance of the feature. 'feature_names': manually list all the feature names of the model, which correspond to the various features in our previous model. 'np.argsort(importances)[::-1]': Arrange the feature importance in descending order and get the sorted feature index. Refer to Table 2, which illustrates the importance of different input variables (ie, feature dimensions). It is worth noting that this is only exemplary and should not limit the output value of this embodiment in actual application.

[0165] feature importance Energy structure 0.3499 Growth cycle 0.1756 Average annual precipitation 0.1323 Raw material transportation distance 0.0923 Unit output 0.0888 Crop Type 0.0675 Average annual sunshine 0.0654 Location 0.0282

[0166] 4. Model prediction and result analysis

[0167] Model prediction: By inputting new data, such as a certain crop, a specific geographical location, climate conditions, raw material transportation distance, and energy structure, the model can output the predicted results of the water footprint. For example, assuming that the crop growth cycle of a certain biomass raw material is 203 days, the unit area yield is 7.5 tons / hectare, the raw material transportation distance is 500 kilometers, the average annual precipitation is 205 mm, the average annual sunshine hours are 2789.7 hours, the proportion of coal power in the energy structure is 70%, the proportion of natural gas is 10%, and the proportion of renewable energy is 20%, then these conditions can be entered and predicted.

[0168]

[0169] Results visualization: Visualization tools are used to display the prediction results of the model, such as the comparison between the actual water footprint and the predicted value, to help users intuitively understand the accuracy of the model. In addition, feature importance charts can be used to show the impact of different input variables on the water footprint.

[0170] 5. Application scenarios and benefit analysis

[0171] The water footprint assessment method provided in this embodiment can help enterprises accurately predict the water footprint of sustainable aviation fuel according to different conditions. Through feature importance analysis, enterprises can optimize planting planning, give priority to planting crops with low water footprint, and reasonably plan transportation routes to reduce resource waste. In addition, the adjustment of energy structure can also be reasonably optimized through model analysis to minimize water resource consumption.

[0172] Based on the description of the application scenario, the water footprint assessment method provided in this embodiment has the following advantages:

[0173] ●High-precision prediction: Compared with traditional linear models, random forests can capture complex nonlinear relationships and improve the accuracy of water footprint prediction.

[0174] ● Strong scalability: The model can handle multi-dimensional variables and is suitable for water footprint assessment under different crops and different geographical conditions. If you need to add other input objects, you can use them as input variables of the data set, which is convenient for subsequent improvement and modification.

[0175] ●Flexibility: The model can adjust inputs according to different crops, production sites and energy structures, and is suitable for a variety of production environments.

[0176] ●Optimize resource management: Through feature importance analysis, the model can help companies identify the factors that have the greatest impact on water consumption, thereby optimizing production processes and reducing water footprint.

[0177] It should be noted that in some of the processes described in the above embodiments and the accompanying drawings, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel, and the sequence numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the sequence numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different application terminals, messages, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0178] Figure 5 A schematic diagram of a computing device provided as another exemplary embodiment of the present application is shown in FIG. Figure 5 As shown, the computing device includes: a memory 50 , a processor 51 and a communication component 52 .

[0179] The processor 51 is coupled to the memory 50 and the communication component 52, and is used to execute the computer program in the memory 50, so as to:

[0180] In response to a water footprint assessment instruction initiated for sustainable aviation fuel to be produced, a feature parameter set to be processed is obtained, wherein the feature parameter set includes attribute feature parameters of raw material crops, geographical environment feature parameters of raw material crops, and / or production environment feature parameters of sustainable aviation fuel;

[0181] Inputting the characteristic parameter set into a pre-trained water footprint assessment model, wherein the water footprint assessment model adopts a random forest regression model, wherein the water footprint assessment model includes a plurality of preset decision trees, and the water footprint assessment model is used to perform water footprint assessment on the characteristic parameter set respectively through the plurality of decision trees;

[0182] In the water footprint assessment model, the characteristic parameter sets are respectively input into the multiple decision trees, so as to generate assessment results respectively through the multiple decision trees;

[0183] The evaluation results generated by the multiple decision trees are aggregated to obtain a final evaluation result as the water footprint evaluation result corresponding to the sustainable aviation fuel.

[0184] In an optional embodiment, when the processor 51 inputs the characteristic parameter set into the pre-trained water footprint assessment model, it can be specifically used to:

[0185] Dividing the characteristic parameters contained in the characteristic parameter set into at least one parameter group, where different parameter groups correspond to different variable types;

[0186] If there is a first parameter group whose variable type is a categorical variable, one-hot encoding is performed on the feature parameters in the first parameter group to obtain corresponding preprocessed parameters;

[0187] If there is a second parameter group whose variable type is a numerical variable, standardizing the characteristic parameters in the second parameter group to obtain corresponding preprocessed parameters;

[0188] The preprocessed characteristic parameter set is input into the water footprint assessment model.

[0189] In an optional embodiment, the attribute parameters of the raw material crops include parameter values ​​under one or more characteristic dimensions of crop category, growth cycle, unit yield and reference water footprint; the geographical environment parameters of the raw material crops include parameter values ​​under one or more characteristic dimensions of geographical location, precipitation and sunshine duration; the production environment parameters of the sustainable aviation fuel include parameter values ​​under one or more characteristic dimensions of transportation distance and energy proportion; wherein, the first parameter group includes parameter values ​​under the crop category and / or the geographical location; and the second parameter group includes parameter values ​​under the growth cycle, the unit yield, the reference water footprint, the precipitation, the sunshine duration, the transportation distance and / or the energy proportion.

[0190] In an optional embodiment, the processor 51 may also be configured to:

[0191] In the process of performing water footprint assessment on the characteristic parameter set by the multiple decision trees in the water footprint assessment model respectively, determining the impact of the assessment result corresponding to each node included in the multiple decision trees, each node corresponding to a characteristic dimension;

[0192] Based on the impact of the evaluation results determined for each node, the feature importance is statistically analyzed in units of feature dimensions;

[0193] Among them, in the multiple decision trees, there are one or more nodes corresponding to the same feature dimension. If there are multiple nodes, the influence of the evaluation results corresponding to the multiple nodes is fused to calculate the feature importance corresponding to the feature dimension.

[0194] In an optional embodiment, the processor 51 may also be configured to:

[0195] Sort the feature dimensions according to the feature importance calculated for different feature dimensions;

[0196] The feature importance of the sorted feature dimensions is visualized to show the contribution of different feature dimensions to the water footprint assessment results.

[0197] In an optional embodiment, the processor 51, during the process of training the water footprint assessment model, may be specifically configured to:

[0198] Under the preset evaluation dimensions, characteristic dimensions are respectively determined to obtain a characteristic dimension set, wherein the evaluation dimensions include one or more factors in the attributes of the raw material crop, the geographical environment of the raw material crop, and / or the production environment of the sustainable aviation fuel;

[0199] For any decision tree to be constructed, select some feature dimensions from the feature dimension set;

[0200] Extract multiple training samples from the training sample set as a training set;

[0201] Based on the selected feature dimensions, construct the decision tree, wherein the feature dimensions serve as nodes in the decision tree;

[0202] Continue to construct other decision trees to form the water footprint assessment model based on the constructed multiple decision trees;

[0203] Based on the training set, the model parameters of the water footprint assessment model are optimized and the decision tree structure is adjusted to complete the training of the water footprint assessment model.

[0204] In an optional embodiment, the processor 51 may also be configured to:

[0205] Extracting multiple training samples from the training sample set as a test set;

[0206] Inputting the test set into the water footprint assessment model to obtain a test result;

[0207] Performing performance verification on the water footprint assessment model in a cross-validation manner to optimize the hyperparameters corresponding to the water footprint assessment model;

[0208] The validation indicators used in the cross-validation process include mean square error and / or mean absolute error.

[0209] In an optional embodiment, when the processor 51 aggregates the evaluation results generated by the plurality of decision trees to obtain a final evaluation result, it may be specifically used to:

[0210] Calculating the mean of the evaluations generated by the plurality of decision trees;

[0211] The calculated mean is taken as the final evaluation result.

[0212] In an optional embodiment, the processor 51 may also be configured to:

[0213] The water footprint assessment results corresponding to the sustainable aviation fuel are visualized.

[0214] Further, if Figure 5 As shown, the computing device also includes: a display 53, a power component 54, an audio component 55 and other components. Figure 5 Only some components are shown schematically, and it does not mean that the computing device only includes Figure 5 Components shown.

[0215] It is worth noting that the technical details in the above-mentioned embodiments of the computing device can refer to the relevant description in the aforementioned method embodiment. In order to save space, they will not be repeated here, but this should not cause a loss in the scope of protection of this application.

[0216] Accordingly, an embodiment of the present application further provides a computer-readable storage medium storing a computer program, which can implement the steps in the above method embodiment when the computer program is executed.

[0217] Accordingly, an embodiment of the present application also provides a computer program product, which can implement the steps in the above method embodiment when the computer program included in the product is executed.

[0218] The above-mentioned memory is used to store computer programs and can be configured to store various other data to support operations on the computing platform. Examples of these data include instructions for any application or method operating on the computing platform, contact data, phone book data, messages, pictures, videos, etc. The memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0219] The above-mentioned communication component is configured to facilitate wired or wireless communication between the device where the communication component is located and other devices. The device where the communication component is located can access a wireless network based on a communication standard, such as WiFi, 2G, 3G, 4G / LTE, 5G and other mobile communication networks, or a combination thereof. In an exemplary embodiment, the communication component receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

[0220] The above-mentioned display includes a screen, and the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation.

[0221] The power supply assembly described above provides power to various components of the device in which the power supply assembly is located. The power supply assembly may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the device in which the power supply assembly is located.

[0222] The above-mentioned audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (MIC), and when the device where the audio component is located is in an operating mode, such as a call mode, a recording mode, and a speech recognition mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in a memory or sent via a communication component. In some embodiments, the audio component also includes a speaker for outputting an audio signal.

[0223] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0224] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0225] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0226] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0227] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0228] The above is only the embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A water footprint assessment method, characterized in that: include: In response to a water footprint assessment instruction initiated for sustainable aviation fuel to be produced, a feature parameter set to be processed is obtained, wherein the feature parameter set includes attribute feature parameters of raw material crops, geographical environment feature parameters of raw material crops, and / or production environment feature parameters of sustainable aviation fuel; Inputting the characteristic parameter set into a pre-trained water footprint assessment model, wherein the water footprint assessment model adopts a random forest regression model, wherein the water footprint assessment model includes a plurality of preset decision trees, and the water footprint assessment model is used to perform water footprint assessment on the characteristic parameter set respectively through the plurality of decision trees; In the water footprint assessment model, the characteristic parameter sets are respectively input into the multiple decision trees, so as to generate assessment results respectively through the multiple decision trees; The evaluation results generated by the multiple decision trees are aggregated to obtain a final evaluation result as the water footprint evaluation result corresponding to the sustainable aviation fuel.

2. The method according to claim 1, characterized in that The characteristic parameter set is input into the pre-trained water footprint assessment model, including: Dividing the characteristic parameters contained in the characteristic parameter set into at least one parameter group, where different parameter groups correspond to different variable types; If there is a first parameter group whose variable type is a categorical variable, one-hot encoding is performed on the feature parameters in the first parameter group to obtain corresponding preprocessed parameters; If there is a second parameter group whose variable type is a numerical variable, standardizing the characteristic parameters in the second parameter group to obtain corresponding preprocessed parameters; The preprocessed characteristic parameter set is input into the water footprint assessment model.

3. The method according to claim 2, characterized in that The attribute parameters of the raw material crops include parameter values ​​under one or more characteristic dimensions of crop category, growth cycle, unit yield and reference water footprint; the geographical environment parameters of the raw material crops include parameter values ​​under one or more characteristic dimensions of geographical location, precipitation and sunshine duration; the production environment parameters of the sustainable aviation fuel include parameter values ​​under one or more characteristic dimensions of raw material transportation distance and energy proportion; wherein, the first parameter group includes parameter values ​​under the crop category and / or the geographical location; and the second parameter group includes parameter values ​​under the growth cycle, the unit yield, the reference water footprint, the precipitation, the sunshine duration, the transportation distance and / or the energy proportion.

4. The method according to claim 1, characterized in that: Also includes: In the process of performing water footprint assessment on the characteristic parameter set by the multiple decision trees in the water footprint assessment model respectively, determining the impact of the assessment result corresponding to each node included in the multiple decision trees, each node corresponding to a characteristic dimension; Based on the impact of the evaluation results determined for each node, the feature importance is statistically analyzed in units of feature dimensions; Among them, in the multiple decision trees, there are one or more nodes corresponding to the same feature dimension. If there are multiple nodes, the influence of the evaluation results corresponding to the multiple nodes is fused to calculate the feature importance corresponding to the feature dimension.

5. The method according to claim 4, characterized in that Also includes: Sort the feature dimensions according to the feature importance calculated for different feature dimensions; The feature importance of the sorted feature dimensions is visualized to show the contribution of different feature dimensions to the water footprint assessment results.

6. The method according to claim 1, characterized in that The training process of the water footprint assessment model includes: Under the preset evaluation dimensions, characteristic dimensions are determined respectively to obtain a characteristic dimension set, wherein the evaluation dimensions include one or more dimensions of attributes of the raw material crop, the geographical environment of the raw material crop, and / or the production environment of the sustainable aviation fuel; For any decision tree to be constructed, select some feature dimensions from the feature dimension set; Extract multiple training samples from the training sample set as a training set; Based on the selected feature dimensions, construct the decision tree, wherein the feature dimensions serve as nodes in the decision tree; Continue to construct other decision trees to form the water footprint assessment model based on the constructed multiple decision trees; Based on the training set, the model parameters of the water footprint assessment model are optimized and the decision tree structure is adjusted to complete the training of the water footprint assessment model.

7. The method according to claim 6, characterized in that Also includes: Extracting multiple training samples from the training sample set as a test set; Inputting the test set into the water footprint assessment model to obtain a test result; Performing performance verification on the water footprint assessment model in a cross-validation manner to optimize the hyperparameters corresponding to the water footprint assessment model; The validation indicators used in the cross-validation process include mean square error and / or mean absolute error.

8. The method according to claim 1, characterized in that Aggregate the evaluation results generated by the multiple decision trees to obtain a final evaluation result, including: Calculating the mean of the evaluations generated by the plurality of decision trees; The calculated mean is taken as the final evaluation result.

9. The method according to claim 1, characterized in that: Also includes: The water footprint assessment results corresponding to the sustainable aviation fuel are visualized.

10. A computing device, characterized in that Includes memory, processor, and communication components; The memory is used to store one or more computer instructions; The processor is coupled to the memory and the communication component, and is configured to execute the one or more computer instructions to execute the water footprint assessment method according to any one of claims 1 to 9.

Images