Green carbon footprint prediction method based on machine learning and related device

Abstract

The invention relates to a green carbon footprint prediction method based on machine learning and a related device, and the method comprises the steps: collecting existing product carbon footprint data, including energy consumption data and raw material use data of each life cycle link of a product and carbon emission data of each stage; dividing the collected data into a training set, a test set and a verification set according to the proportion of 8: 1: 1, training an artificial neural network model on the training set, testing the performance of the model on the test set, and finally verifying the effectiveness of the model on the verification set; inputting the training set data into an artificial neural network for deep learning; introducing a mean square error loss function to carry out gradient optimization on the artificial neural network model; and applying the artificial neural network model to carbon footprint prediction calculation of other products. The machine learning algorithm is utilized to predict the carbon footprint of the product, data acquisition can be greatly simplified, and the calculation speed is higher than that of a traditional calculation method.

Description

Technical Field

[0001] The present invention relates to the field of product carbon footprint calculation, and in particular to a green carbon footprint prediction method based on machine learning and a related device. Background Art

[0002] As the global climate change problem becomes increasingly serious, reducing carbon emissions and implementing sustainable development strategies have become the focus of attention of governments and enterprises around the world. my country has successively introduced a series of policies to accelerate the improvement of the carbon footprint management level of key domestic products, promote the green and low-carbon transformation of related industries, and help achieve the goals of carbon peak and carbon neutrality. As an important indicator to measure the impact of human activities on the environment, the green carbon footprint is of great significance for formulating effective carbon emission reduction strategies. Accurately predicting the green carbon footprint will help policymakers, enterprises and individuals better understand the sources of carbon emissions and take corresponding emission reduction measures.

[0003] Traditional carbon footprint calculation methods generally use the Life Cycle Assessment method to calculate the carbon footprint of a product. However, the LCA method has the following shortcomings:

[0004] 1. Difficulty in obtaining data

[0005] Incomplete data: It is difficult to obtain comprehensive and accurate carbon footprint data, especially data involving complex supply chains.

[0006] 2. Complex methods

[0007] Complex calculations: The traditional life cycle assessment (LCA) method has a complex calculation process and requires a lot of time and expertise. Summary of the invention

[0008] In view of this, it is necessary to provide a green carbon footprint prediction method and related devices based on machine learning to solve the inconvenience brought by the traditional life cycle assessment method, simplify the data collection scale and improve the efficiency of carbon footprint calculation.

[0009] The present invention provides a green carbon footprint prediction method based on machine learning, comprising the following steps:

[0010] Collect existing product carbon footprint data, including energy consumption data, raw material usage data and carbon emission data at each stage of the product life cycle;

[0011] Divide the collected data into a training set, a test set and a validation set in a ratio of 8:1:1, wherein the training set is used to train the artificial neural network model, the test set is used to test the performance of the artificial neural network model, and the validation set is used to verify the effectiveness of the artificial neural network;

[0012] Input the training set data into the artificial neural network model for training, wherein the input data includes the energy consumption data and raw material usage data of each life cycle link of the product, and the expected output includes the carbon emission data of each stage of the product life cycle link and the total carbon emission data, wherein the sum of the carbon emissions of each stage is the total carbon emission;

[0013] The mean square error loss function is introduced to perform gradient optimization on the artificial neural network model to obtain the artificial neural network model for testing;

[0014] Verify the effectiveness of the artificial neural network model on the test set, and verify the performance of the artificial neural network model on the validation set;

[0015] The validated artificial neural network model is used in carbon footprint prediction calculations for other products.

[0016] Furthermore, the artificial neural network includes an input layer, a hidden layer and an output layer, wherein the input layer receives the feature vector x, the output layer outputs the final prediction y, and the hidden layer is between the input layer and the output layer;

[0017] The input and output variables of each layer in the artificial neural network are explained as follows:

[0018] x=(x1 x2...x n ): Input vector, w ih : The connection weight between the input layer and the hidden layer;

[0019] h i =(h i1 h i2 ...h ip ): Hidden layer input vector, w oh : The connection weight between the hidden layer and the output layer;

[0020] h o =(h o1 h o2 ...h op ): Hidden layer output vector, b n : The threshold of each neuron in the hidden layer;

[0021] y i =(y i1 h y2 ...h yq ): Output layer input vector, b o : The threshold of each neuron in the output layer;

[0022] y o =(y o1 y o2 ...y oq): Output layer output vector, k = 1, 2...m: Sample data and number:

[0023] d o =(d1 d2...d q ): expected output vector, f: activation function.

[0024] Furthermore, the step of inputting the training set data into the artificial neural network model for training specifically includes:

[0025] Initialize the weights of the artificial neural network and assign a random number in the interval (-1, 1) to each connection weight;

[0026] The training set data x(k) and the expected output d o (k) inputting the artificial neural network model;

[0027] The x(k) includes energy consumption data and raw material usage data of each life cycle link of the product;

[0028] The d o (k) including carbon emission data at each stage of the product life cycle and total carbon emission data, where the sum of carbon emissions at each stage is the total carbon emission;

[0029] Train an artificial neural network, where the input and output of each neuron in the hidden layer and output layer are:

[0030]

[0031] Furthermore, the introduction of the mean square error loss function to perform gradient optimization on the artificial neural network model specifically includes:

[0032] The selected mean square error loss function is:

[0033] Where: L(θ) is the mean square error loss function, which depends on the model parameters θ; y i is the ith true value, is the i-th predicted value, N is the number of samples;

[0034] Parameter initialization: Set the predicted value to Initialize θ0 and θ1;

[0035] Perform forward propagation: Calculate the predicted value for each sample

[0036] Calculate the loss: Calculate the mean square error of all samples:

[0037]

[0038] Compute the gradient:

[0039]

[0040] Update parameters:

[0041]

[0042] Among them, α is the learning rate;

[0043] Iterative optimization: Repeat the above steps until the stopping condition is met.

[0044] A green carbon footprint prediction device based on machine learning, comprising:

[0045] Carbon footprint data collection module, used to collect existing product carbon footprint data, including energy consumption data, raw material usage data and carbon emission data at each stage of the product life cycle;

[0046] A data set segmentation module is used to divide the collected data into a training set, a test set and a validation set in a ratio of 8:1:1, wherein the training set is used to train the artificial neural network model, the test set is used to test the performance of the artificial neural network model, and the validation set is used to verify the effectiveness of the artificial neural network;

[0047] A model training module is used to input the training set data into the artificial neural network model for training, wherein the input data includes the energy consumption data and raw material usage data of each life cycle link of the product, and the expected output includes the carbon emission data of each stage of the product life cycle link and the total carbon emission data, wherein the sum of the carbon emissions of each stage is the total carbon emission;

[0048] A model optimization module is used to introduce a mean square error loss function to perform gradient optimization on the artificial neural network model to obtain an artificial neural network model for testing;

[0049] The model verification module is used to verify the effectiveness of the artificial neural network model on the test set and to verify the performance of the artificial neural network model on the verification set;

[0050] The carbon footprint prediction module is used to apply the verified artificial neural network model to the carbon footprint prediction calculation of other products.

[0051] Furthermore, the artificial neural network includes an input layer, a hidden layer and an output layer, wherein the input layer receives the feature vector x, the output layer outputs the final prediction y, and the hidden layer is between the input layer and the output layer;

[0052] The input and output variables of each layer in the artificial neural network are explained as follows:

[0053] x=(x1 x2...x n): Input vector, w ih : The connection weight between the input layer and the hidden layer;

[0054] h i =(h i1 h i2 ...h ip ): Hidden layer input vector, w oh : The connection weight between the hidden layer and the output layer;

[0055] h o =(h o1 h o2 ...h op ): Hidden layer output vector, b n : The threshold of each neuron in the hidden layer;

[0056] y i =(y i1 h y2 ...h yq ): Output layer input vector, b o : The threshold of each neuron in the output layer;

[0057] y o =(y o1 y o2 ...y oq ): Output layer output vector, k = 1, 2...m: Sample data and number:

[0058] d o =(d1 d2...d q ): expected output vector, f: activation function.

[0059] Furthermore, the model training module is specifically used for:

[0060] Initialize the weights of the artificial neural network and assign a random number in the interval (-1, 1) to each connection weight;

[0061] The training set data x(k) and the expected output d o (k) inputting the artificial neural network model;

[0062] The x(k) includes energy consumption data and raw material usage data of each life cycle link of the product;

[0063] The d o (k) including carbon emission data at each stage of the product life cycle and total carbon emission data, where the sum of carbon emissions at each stage is the total carbon emission;

[0064] Train an artificial neural network, where the input and output of each neuron in the hidden layer and output layer are:

[0065]

[0066] 9. Furthermore, the model optimization module is specifically used for:

[0067] The selected mean square error loss function is:

[0068] Where: L(θ) is the mean square error loss function, which depends on the model parameters θ; y i is the ith true value, is the i-th predicted value, N is the number of samples;

[0069] Parameter initialization: Set the predicted value to Initialize θ0 and θ1;

[0070] Perform forward propagation: Calculate the predicted value for each sample

[0071] Calculate the loss: Calculate the mean square error of all samples:

[0072]

[0073] Compute the gradient:

[0074]

[0075] Update parameters:

[0076]

[0077] Among them, α is the learning rate;

[0078] Iterative optimization: Repeat the above steps until the stopping condition is met.

[0079] A green carbon footprint prediction system based on machine learning, comprising: a computer-readable storage medium and a processor;

[0080] The computer-readable storage medium is used to store executable instructions;

[0081] The processor is used to read the executable instructions stored in the computer-readable storage medium to execute the green carbon footprint prediction method based on machine learning.

[0082] A non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the green carbon footprint prediction method based on machine learning.

[0083] The present invention has the following beneficial effects:

[0084] 1) It is difficult to obtain comprehensive and accurate carbon footprint data based on traditional life cycle assessment methods, especially data involving complex supply chains. The present invention uses machine learning algorithms to predict product carbon footprints, which greatly simplifies data collection and complex calculation processes;

[0085] 2) The present invention uses a machine learning algorithm to predict the carbon footprint of a product. Compared with traditional calculation methods, this method greatly improves the calculation time and efficiency of the carbon footprint of a product;

[0086] 3) The present invention can be extended to the carbon footprint calculation of products in other fields and has good generalization performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] Figure 1 A schematic diagram of a green carbon footprint prediction method based on machine learning provided in an embodiment of the present invention;

[0088] Figure 2 Schematic diagram of an artificial neural network. DETAILED DESCRIPTION

[0089] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0090] The embodiment of the present invention provides a green carbon footprint prediction method based on machine learning, comprising the following steps:

[0091] Step S1, collect the existing product carbon footprint data calculated based on the life cycle assessment method, including energy consumption data, raw material usage data and carbon emission data of each life cycle link; wherein, the carbon emission data is data strictly calculated by the life cycle assessment method, and the energy consumption data and raw material usage data are the easiest and most direct data to obtain, but not all data that contribute to carbon emissions.

[0092] The carbon emission data calculated by the life cycle assessment method includes all carbon emissions in each life cycle of the product, including indirect carbon emissions and direct carbon emissions. Direct carbon emissions data are easy to obtain, while indirect carbon emissions are difficult to count and data collection and calculation are complex.

[0093] Statistics are collected on energy usage and raw material usage in each stage of the product life cycle, including raw material collection, production, transportation, and disposal, and energy consumption data and raw material usage data that are easily available in each stage of the product life cycle are calculated.

[0094] Among them, the carbon emissions calculated based on the energy consumption data and raw material usage data are smaller than the carbon emissions calculated by the life cycle assessment method, but there is a certain nonlinear relationship between the two.

[0095] The product carbon footprint may be the carbon footprint of any product, and the embodiments of the present invention are mainly aimed at equipment in power systems.

[0096] Collect data from various power equipment. Take the S20-400KVA transformer as an example. The specific data is shown in Table 1:

[0097] Table 1S20-400KVA Transformer Carbon Footprint Calculation Table

[0098]

[0099] Step S2, the collected product carbon footprint data is divided into a training set, a test set and a validation set according to a ratio of 8:1:1, wherein the training set is used to train the artificial neural network model to improve the generalization ability of the model;

[0100] Furthermore, since the materials, energy consumption, and processes in the production stages of each product are different, the input vector is a very high-dimensional vector, in which the dimension values ​​not involved in the product are 0, so that the length of the input vector can be unified.

[0101] The data of the training set, test set and validation set include energy consumption data, raw material usage data and carbon emission data at each stage of the product's life cycle.

[0102] Step S3, inputting the training set data into the artificial neural network model for training;

[0103] The input data includes energy consumption data and raw material usage data of each life cycle link of the product. The expected output includes carbon emission data of each stage of the product life cycle and total carbon emission data, where the sum of carbon emissions of each stage is the total carbon emissions.

[0104] The artificial neural network includes an input layer, a hidden layer and an output layer. Figure 2 As shown in Figure 1, the input layer accepts the feature vector x, the output layer produces the final prediction y, and the hidden layer is between the input layer and the output layer. The values ​​generated in the hidden layer are not as directly visible as the sample matrix x used by the input layer or the label matrix y used by the output layer.

[0105] The input and output variables of each layer in the artificial neural network are explained as follows:

[0106] x=(x1 x2...x n ): Input vector, w ih: The connection weight between the input layer and the hidden layer;

[0107] h i =(h i1 h i2 ...h ip ): Hidden layer input vector, w oh : The connection weight between the hidden layer and the output layer;

[0108] h o =(h o1 h o2 ...h op ): Hidden layer output vector, b n : The threshold of each neuron in the hidden layer;

[0109] y i =(y i1 h y2 ...h yq ): Output layer input vector, b o : The threshold of each neuron in the output layer;

[0110] y o =(y o1 y o2 ...y oq ): Output layer output vector, k = 1, 2...m: Sample data and number:

[0111] d o =(d1 d2...d q ): expected output vector, f: activation function.

[0112] Step S3 specifically includes:

[0113] Initialize the weights of the artificial neural network and assign a random number in the interval (-1, 1) to each connection weight;

[0114] The training set data x(k) and the expected output d o (k) inputting the artificial neural network model;

[0115] The x(k) includes energy consumption data and raw material usage data of each life cycle link of the product;

[0116] The d o (k) including carbon emission data at each stage of the product life cycle and total carbon emission data, where the sum of carbon emissions at each stage is the total carbon emission;

[0117] Train an artificial neural network, where the input and output of each neuron in the hidden layer and output layer are:

[0118]

[0119] Step S4, using mean square error loss function to perform gradient optimization on the artificial neural network model;

[0120] Where: L(θ) is the mean square error loss function, which depends on the model parameters θ; y i is the ith true value, is the i-th predicted value and N is the number of samples.

[0121] Parameter initialization: Set the predicted value to Initialize θ0 and θ1;

[0122] Forward propagation: Calculate the predicted value for each sample

[0123] Calculate the loss: Calculate the mean square error of all samples

[0124]

[0125] Compute the gradient:

[0126]

[0127] Update parameters:

[0128]

[0129] Among them, α is the learning rate;

[0130] Iterative optimization: Repeat the above steps until the stopping condition is met.

[0131] After the above steps, the model is trained and gradient optimized to obtain an artificial neural network model for testing;

[0132] Step S5, verifying the effectiveness of the artificial neural network model on the test set, and verifying the performance of the artificial neural network model on the validation set.

[0133] The carbon emission data of each stage of the test set are shown in Table 2:

[0134] Table 2 Test set carbon footprint data

[0135] Raw material collection stage Transportation Stage Production stage Recycling stage total 2013.231 12.367 48.327 2.534 2076.459

[0136] The results of verifying the effectiveness of the artificial neural network model on the test set are shown in Table 3:

[0137] Table 3. Results of the test set verifying the effectiveness of the artificial neural network model

[0138] Raw material collection stage Transportation Stage Production stage Recycling stage total Original data 2013.231 12.367 48.327 2.534 2076.459 Model Output 1968.364 11.751 46.235 2.465 2028.815 error / % 2.229 4.981 4.329 2.723 2.294

[0139] It can be seen that the trained artificial neural network model performs well on the test set, with the total error controlled at 2.294%, showing good prediction performance;

[0140] The carbon emission data of each stage of the validation set are shown in Table 4:

[0141] Table 4 Validation set carbon footprint data

[0142] Raw material collection stage Transportation Stage Production stage Recycling stage total 3121.831 18.4506 70.189 3.834 3214.305

[0143] The results of verifying the artificial neural network model on the validation set are shown in Table 5:

[0144] Table 5. Results of artificial neural network model verification on validation set

[0145] Raw material collection stage Transportation Stage Production stage Recycling stage total Original data 3121.831 18.4506 70.189 3.834 3214.305 Model Output 3084.367 18.016 76.546 4.023 3182.952 error / % 1.200 2.355 -9.057 -4.930 0.975

[0146] It can be seen that the trained artificial neural network model performs well on the validation set, with the total error controlled at 0.975%, showing good prediction performance;

[0147] Through the above steps, it is proved that the green carbon footprint prediction method based on machine learning proposed in the present invention is reliable and feasible in predicting carbon footprint.

[0148] Step S6, using the trained artificial neural network model for carbon footprint prediction calculation of other products.

[0149] Another aspect of the present invention provides a green carbon footprint prediction device based on machine learning, characterized in that it includes:

[0150] Carbon footprint data collection module, used to collect existing product carbon footprint data, including energy consumption data, raw material usage data and carbon emission data at each stage of the product life cycle;

[0151] A data set segmentation module is used to divide the collected data into a training set, a test set and a validation set in a ratio of 8:1:1, wherein the training set is used to train the artificial neural network model, the test set is used to test the performance of the artificial neural network model, and the validation set is used to verify the effectiveness of the artificial neural network;

[0152] A model training module is used to input the training set data into the artificial neural network model for training, wherein the input data includes the energy consumption data and raw material usage data of each life cycle link of the product, and the expected output includes the carbon emission data of each stage of the product life cycle link and the total carbon emission data, wherein the sum of the carbon emissions of each stage is the total carbon emission;

[0153] A model optimization module is used to introduce a mean square error loss function to perform gradient optimization on the artificial neural network model to obtain an artificial neural network model for testing;

[0154] The model verification module is used to verify the effectiveness of the artificial neural network model on the test set and to verify the performance of the artificial neural network model on the verification set;

[0155] The carbon footprint prediction module is used to apply the verified artificial neural network model to the carbon footprint prediction calculation of other products.

[0156] Another aspect of the present invention provides a green carbon footprint prediction system based on machine learning, comprising: a computer-readable storage medium and a processor;

[0157] The computer-readable storage medium is used to store executable instructions;

[0158] The processor is used to read the executable instructions stored in the computer-readable storage medium to execute the green carbon footprint prediction method based on machine learning described in the first aspect.

[0159] On the other hand, the present invention provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the green carbon footprint prediction method based on machine learning described in the first aspect.

[0160] 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.

[0161] 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.

[0162] 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 including 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.

[0163] 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.

[0164] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A green carbon footprint prediction method based on machine learning, characterized in that: The steps include: Collect existing product carbon footprint data, including energy consumption data, raw material usage data and carbon emission data at each stage of the product life cycle; Divide the collected data into a training set, a test set and a validation set in a ratio of 8:1:1, wherein the training set is used to train the artificial neural network model, the test set is used to test the performance of the artificial neural network model, and the validation set is used to verify the effectiveness of the artificial neural network; Input the training set data into the artificial neural network model for training, wherein the input data includes the energy consumption data and raw material usage data of each life cycle link of the product, and the expected output includes the carbon emission data of each stage of the product life cycle link and the total carbon emission data, wherein the sum of the carbon emissions of each stage is the total carbon emission; The mean square error loss function is introduced to perform gradient optimization on the artificial neural network model to obtain the artificial neural network model for testing; Verify the effectiveness of the artificial neural network model on the test set, and verify the performance of the artificial neural network model on the validation set; The validated artificial neural network model is used in carbon footprint prediction calculations for other products.

2. The green carbon footprint prediction method based on machine learning according to claim 1 is characterized in that: The artificial neural network includes an input layer, a hidden layer and an output layer, wherein the input layer receives the feature vector x, the output layer outputs the final prediction y, and the hidden layer is between the input layer and the output layer; The input and output variables of each layer in the artificial neural network are explained as follows: x=(x1 x2...x n ): Input vector, w ih : The connection weight between the input layer and the hidden layer; h i =(h i1 h i2 ...h ip ): Hidden layer input vector, w oh : The connection weight between the hidden layer and the output layer; h o =(h o1 h o2 ...h op ): Hidden layer output vector, b n : The threshold of each neuron in the hidden layer; y i =(y i1 h y2 ...h yq ): Output layer input vector, b o : The threshold of each neuron in the output layer; y o =(y o1 y o2 ...y oq ): Output layer output vector, k = 1, 2…m: Sample data and number: d o =(d1 d2...d q ): expected output vector, f: activation function.

3. The green carbon footprint prediction method based on machine learning according to claim 2 is characterized in that: The step of inputting the training set data into the artificial neural network model for training specifically includes: Initialize the weights of the artificial neural network and assign a random number in the interval (-1, 1) to each connection weight; The training set data x(k) and the expected output d o (k) inputting the artificial neural network model; The x(k) includes energy consumption data and raw material usage data of each life cycle link of the product; The d o (k) including carbon emission data at each stage of the product life cycle and total carbon emission data, where the sum of carbon emissions at each stage is the total carbon emission; Train an artificial neural network, where the input and output of each neuron in the hidden layer and output layer are: ho h (k)=f(hi h (k)),h=1,2,...,p; o=1,2,...,q,yo o (k) = f(yi o (k)), o = 1, 2, ..., q.

4. The green carbon footprint prediction method based on machine learning according to claim 3 is characterized in that: The introduction of the mean square error loss function to perform gradient optimization on the artificial neural network model specifically includes: The selected mean square error loss function is: Where: L(θ) is the mean square error loss function, which depends on the model parameters θ; y i is the ith true value, is the i-th predicted value, N is the number of samples; Parameter initialization: Set the predicted value to Initialize θ0 and θ1; Perform forward propagation: Calculate the predicted value for each sample Calculate the loss: Calculate the mean square error of all samples: Compute the gradient: Update parameters: Among them, α is the learning rate; Iterative optimization: Repeat the above steps until the stopping condition is met.

5. A green carbon footprint prediction device based on machine learning, characterized in that: include: Carbon footprint data collection module, used to collect existing product carbon footprint data, including energy consumption data, raw material usage data and carbon emission data at each stage of the product life cycle; A data set segmentation module is used to divide the collected data into a training set, a test set and a validation set in a ratio of 8:1:1, wherein the training set is used to train the artificial neural network model, the test set is used to test the performance of the artificial neural network model, and the validation set is used to verify the effectiveness of the artificial neural network; A model training module is used to input the training set data into the artificial neural network model for training, wherein the input data includes the energy consumption data and raw material usage data of each life cycle link of the product, and the expected output includes the carbon emission data of each stage of the product life cycle link and the total carbon emission data, wherein the sum of the carbon emissions of each stage is the total carbon emission; A model optimization module is used to introduce a mean square error loss function to perform gradient optimization on the artificial neural network model to obtain an artificial neural network model for testing; The model verification module is used to verify the effectiveness of the artificial neural network model on the test set and to verify the performance of the artificial neural network model on the verification set; The carbon footprint prediction module is used to apply the verified artificial neural network model to the carbon footprint prediction calculation of other products.

6. The green carbon footprint prediction device based on machine learning according to claim 5, characterized in that: The artificial neural network includes an input layer, a hidden layer and an output layer, wherein the input layer receives the feature vector x, the output layer outputs the final prediction y, and the hidden layer is between the input layer and the output layer; The input and output variables of each layer in the artificial neural network are explained as follows: x=(x1 x2...x n ): Input vector, w ih : The connection weight between the input layer and the hidden layer; h i =(h i1 h i2 ...h ip ): Hidden layer input vector, w oh : The connection weight between the hidden layer and the output layer; h o =(h o1 h o2 ...h op ): Hidden layer output vector, b n : The threshold of each neuron in the hidden layer; y i =(y i1 h y2 ...h yq ): Output layer input vector, b o : The threshold of each neuron in the output layer; y o =(y o1 y o2 ...y oq ): Output layer output vector, k = 1, 2...m: Sample data and number: d o =(d1 d2...d q ): expected output vector, f: activation function.

7. The green carbon footprint prediction device based on machine learning according to claim 6, characterized in that: The model training module is specifically used for: Initialize the weights of the artificial neural network and assign a random number in the interval (-1, 1) to each connection weight; The training set data x(k) and the expected output d o (k) inputting the artificial neural network model; The x(k) includes energy consumption data and raw material usage data of each life cycle link of the product; The d o (k) including carbon emission data at each stage of the product life cycle and total carbon emission data, where the sum of carbon emissions at each stage is the total carbon emission; Train an artificial neural network, where the input and output of each neuron in the hidden layer and output layer are: ho h (k)=f(hi h (k)),h=1,2,...,p; yo o (k) = f(yi o (k)), o = 1, 2, ..., q.

8. The green carbon footprint prediction device based on machine learning according to claim 7, characterized in that: The model optimization module is specifically used for: The selected mean square error loss function is: Where: L(θ) is the mean square error loss function, which depends on the model parameters θ; y i is the ith true value, is the i-th predicted value, N is the number of samples; Parameter initialization: Set the predicted value to Initialize θ0 and θ1; Perform forward propagation: Calculate the predicted value for each sample Calculate the loss: Calculate the mean square error of all samples: Compute the gradient: Update parameters: Among them, α is the learning rate; Iterative optimization: Repeat the above steps until the stopping condition is met.

9. A green carbon footprint prediction system based on machine learning, comprising: A computer readable storage medium and a processor; The computer-readable storage medium is used to store executable instructions; The processor is used to read the executable instructions stored in the computer-readable storage medium to execute the green carbon footprint prediction method based on machine learning described in any one of claims 1-4.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the green carbon footprint prediction method based on machine learning as described in any one of claims 1 to 4.

Images