A data collection and analysis method for detecting polymer water-absorbing materials

Abstract

The invention discloses a data collection and analysis method for detecting polymer water-absorbing materials, and relates to the field of polymer water-absorbing materials. The method comprises: collecting water absorption performance data of polymer water-absorbing materials under different environmental conditions; analyzing the water absorption performance data, obtaining key performance indicators, and constructing a prediction model to screen out several environmental condition attributes that have the greatest impact on the water absorption performance; retraining the prediction model based on the obtained key performance indicators and several environmental condition attributes; using the retrained prediction model to predict the water absorption performance of the polymer water-absorbing material, and optimizing the formula or preparation process of the polymer water-absorbing material according to the prediction result. The invention predicts the water absorption performance of the polymer water-absorbing material and shortens the running time; at the same time, the prediction model is retrained by concentrating the data of key performance indicators and environmental condition attributes to improve the prediction accuracy and generalization ability of the model.

Description

Technical Field

[0001] The invention relates to the field of polymer water-absorbing materials, and in particular to a data collection and analysis method for detecting polymer water-absorbing materials. Background Art

[0002] Polymer absorbent materials, such as super absorbent polymers (SAP), are widely used in agriculture, sanitary products, medical treatment, construction and other fields due to their extremely high water absorption and water retention capacity. Such materials can usually absorb and retain water tens or even hundreds of times their own weight. With the continuous increase in application demand, higher requirements are placed on the performance testing of polymer absorbent materials.

[0003] Traditional performance testing methods for polymer water-absorbing materials mainly rely on weight increase method, centrifugation method, etc. Although these methods can provide basic information on water absorption rate and water retention capacity, there are some deficiencies in the process of data collection and analysis. For example, data collection efficiency is low, real-time monitoring cannot be performed, and it is difficult to fully utilize structure and performance information for comprehensive analysis.

[0004] There is no data collection and analysis method for detecting polymer water-absorbing materials in the prior art. General data collection and analysis methods, such as Chinese patent 201711142218.0, disclose an intelligent data collection and analysis method and system, which realizes data collection and can discard invalid data when the wind direction is wrong, discard abnormal fluctuation data, and restore accurate data of soluble components in rainfall, eliminating the adverse effects of wind direction and rainfall on data collection. However, the above-mentioned data collection and analysis method cannot be applied to the detection of polymer water-absorbing materials, and the accuracy is often not high when predicting material properties, which is difficult to meet the high-standard industrial application requirements. At the same time, a large number of experiments are required to determine the optimal formula or process parameters, which is not only time-consuming but also costly.

[0005] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0006] In view of the problems in the related art, the present invention proposes a data collection and analysis method for detecting polymer water-absorbing materials to overcome the above technical problems existing in the existing related art.

[0007] To this end, the specific technical solution adopted by the present invention is as follows:

[0008] A data collection and analysis method for detecting a polymer water-absorbing material, the data collection and analysis method for detecting a polymer water-absorbing material comprises the following steps:

[0009] S1. Collect water absorption performance data of polymer water-absorbing materials under different environmental conditions.

[0010] S2. Analyze the water absorption performance data, obtain key performance indicators, and build a prediction model to screen out several environmental condition attributes that have the greatest impact on water absorption performance.

[0011] S3. Retrain the prediction model based on the obtained key performance indicators and several environmental condition attributes.

[0012] S4. Use the retrained prediction model to predict the water absorption performance of the polymer water absorbent material, and optimize the formula or preparation process of the polymer water absorbent material based on the prediction results.

[0013] Furthermore, in order to predict the water absorption properties of polymer water-absorbing materials and improve the prediction accuracy and generalization ability of the model:

[0014] Analyzing the water absorption performance data, obtaining key performance indicators, and building a prediction model to screen out several environmental condition attributes that have the greatest impact on water absorption performance includes the following steps:

[0015] S21, quantifying and preprocessing the non-numerical features in the water absorption performance data and the environmental condition data;

[0016] S22, evaluate the weight and attribute volatility of environmental condition data to obtain key performance indicators;

[0017] S23, constructing several regression decision subtrees based on the random forest regression analysis algorithm, calculating the out-of-bag error, and constructing a prediction model;

[0018] S24, using the out-of-bag error and the running time of the prediction model as indicators to perform parameter optimization;

[0019] S25. Obtain several environmental condition attributes that have the greatest impact on water absorption performance.

[0020] Further, evaluating the weight and attribute volatility of environmental condition data and obtaining key performance indicators includes the following steps:

[0021] S221, obtaining the frequency of each environmental condition attribute by counting the number of times each environmental condition attribute appears in all environmental condition data, and comparing the number of times each environmental condition attribute appears with the total number of times all environmental condition attributes appear to obtain the weight of each environmental condition attribute;

[0022] S222. Analyze the volatility of each environmental condition attribute over time;

[0023] S223, obtaining a comprehensive score of each annular condition attribute by calculating the product of the volatility of the environmental condition attribute and the weight of the environmental condition attribute;

[0024] S224. Select several environmental condition attributes with the highest comprehensive scores as key performance indicators.

[0025] Furthermore, based on the random forest regression analysis algorithm, several regression decision subtrees are constructed and the out-of-bag error is calculated. The construction of the prediction model includes the following steps:

[0026] S231, using the environmental condition data as a feature set and the water absorption performance data as a label set, and forming a data set through the feature set and the label set;

[0027] S232. Using the bagging idea and random subspace idea in random forest, several sample subsets and feature subspaces are randomly generated, and several regression decision subtrees are constructed;

[0028] S233, training each regression decision subtree through the sample subset until the stopping condition is met, and for each split node, searching for the best split point through the feature subspace;

[0029] S234, calculating the out-of-bag error of each regression decision subtree, and taking the average of the out-of-bag errors of all regression decision subtrees as the overall out-of-bag error of the prediction model;

[0030] S235. All regression decision subtrees are combined to form a prediction model, and the generalization ability of the prediction model is evaluated through an independent test set.

[0031] Furthermore, the parameter optimization using the out-of-bag error and the prediction model running time as indicators includes the following steps:

[0032] S241, select the number of features m when the initial split node is selected, and fix the value of m, obtain the change of the out-of-bag error and the prediction model training time with the number of regression decision subtrees, and find the number of regression decision subtrees that minimizes the out-of-bag error;

[0033] S242, select the number of regression decision subtrees obtained in S241, obtain the change of out-of-bag error and prediction model training time with the number of features when splitting the node, and find the number of features when splitting the node that minimizes the out-of-bag error;

[0034] S243, repeat S241-S242 until the out-of-bag error no longer decreases or reaches a preset threshold.

[0035] Further, obtaining several environmental condition attributes that have the greatest impact on water absorption performance includes the following steps:

[0036] S251. Select several environmental condition attributes with the highest scores from the feature importance evaluation;

[0037] S252. Use charts to visualize the importance scores of various environmental condition attributes.

[0038] Furthermore, based on the obtained key performance indicators and several environmental condition attributes, retraining the prediction model includes the following steps:

[0039] S31, obtaining key performance indicators and several environmental condition attributes with the highest scores, and obtaining the final environmental condition selection attributes;

[0040] S32. Use the data of the final environmental condition selection attributes to retrain the prediction model.

[0041] Furthermore, in order to systematically optimize the material formula or preparation process, find the optimal high-analysis water-absorbing material formula or process parameters to maximize performance and reduce costs:

[0042] According to the prediction results, optimizing the formula or preparation process of the polymer water-absorbing material includes the following steps:

[0043] S41, constructing a parameter optimization model according to the prediction results of the prediction model and a combination of several randomly selected recipes or preparation process parameters;

[0044] S42, performing iterative optimization based on the acquisition function, and updating the parameter optimization model;

[0045] S43. Select the formula or process parameters with the best water absorption performance predicted by the parameter optimization model as the final optimization result.

[0046] Furthermore, according to the prediction results of the prediction model and a combination of several randomly selected recipes or preparation process parameters, constructing a parameter optimization model includes the following steps:

[0047] S411. Construct an objective function, and the objective function is:

[0048]

[0049] In the formula, f ( x ) represents the objective function value, x A vector representing all recipe or process parameters;

[0050] SR ( x ) means according to the parameter x Predicted water absorption rate, MWC ( x ) means according to the parameter x Predicted maximum water absorption, WR ( x ) means according to the parameter x predicted water retention capacity;

[0051] w 1 , w 2 , w 3 All represent weight coefficients;

[0052] S412, randomly selecting a combination of several formulas or process parameters, and using the prediction model to predict, to obtain data on water absorption performance, and constructing an initial data set based on the randomly selected combinations of several formulas or process parameters and the obtained data on water absorption performance;

[0053] S413. Establish a parameter optimization model using the initial data set, and predict the water absorption performance of a polymer water-absorbing material with an unknown formula or process parameter combination through the parameter optimization model.

[0054] Further, performing iterative optimization based on the acquisition function and updating the parameter optimization model include the following steps:

[0055] S421, selecting an acquisition function, and determining the next formula or process parameter combination for improving water absorption performance through the acquisition function;

[0056] S422, using the acquisition function to select the next experimental scheme, and adding the new water absorption performance data to the initial data set to form an experimental data set;

[0057] S423: Use the experimental data set to update the parameter optimization model, and repeat the updating of the parameter optimization model until a preset number of iterations is reached.

[0058] The beneficial effects of the present invention are:

[0059] (1) The present invention provides a data collection and analysis method for detecting polymer water-absorbing materials. A prediction model is constructed based on a random forest regression algorithm to predict the water absorption performance of polymer water-absorbing materials. The parameters are optimized by using the out-of-bag error and the running time of the prediction model as indicators to improve the performance of the prediction model and shorten the running time. The weight and volatility of environmental conditions are evaluated to identify and quantify the contribution of each environmental condition to the performance, thereby helping to focus on the most critical factors and determine the environmental factors that have the greatest impact on the water absorption performance. The prediction model is retrained by focusing on the data of key performance indicators and environmental condition attributes to improve the prediction accuracy and generalization ability of the model.

[0060] (2) The present invention constructs a parameter optimization model based on the prediction results of the prediction model and a combination of several randomly selected formulas or preparation process parameters, systematically optimizes the material formula or preparation process, finds the optimal high-analysis water-absorbing material formula or process parameters, maximizes performance, and reduces costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0062] Figure 1 The present invention is a flowchart of a data collection and analysis method for detecting polymer water-absorbing materials according to an embodiment of the present invention. DETAILED DESCRIPTION

[0063] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in the field should be able to understand other possible implementations and advantages of the present invention. The components in the figures are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0064] According to an embodiment of the present invention, a data collection and analysis method for detecting polymer water-absorbing materials is provided.

[0065] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, according to the data collection and analysis method for detecting polymer water-absorbing materials according to an embodiment of the present invention, the data collection and analysis method for detecting polymer water-absorbing materials comprises the following steps:

[0066] S1. Collect water absorption performance data of polymer water-absorbing materials under different environmental conditions.

[0067] First, clarify the water absorption performance parameters that need to be tested, such as water absorption rate, maximum water absorption and water retention capacity. Determine the environmental condition parameters, such as temperature, humidity, pH value, ion concentration, etc.

[0068] Select or prepare a series of samples of polymeric water-absorbing materials. Make sure the size, shape and weight of the samples are standardized so that the test results are comparable.

[0069] S2. Analyze the water absorption performance data, obtain key performance indicators, and build a prediction model to screen out several environmental condition attributes that have the greatest impact on water absorption performance;

[0070] In a further embodiment, analyzing the water absorption performance data, obtaining key performance indicators, and constructing a prediction model to screen out several environmental condition attributes that have the greatest impact on the water absorption performance include the following steps:

[0071] S21. Quantify the non-numerical features in the water absorption performance data and environmental condition data, and perform preprocessing (including processing missing values, outlier detection and correction, normalization or standardization of numerical data, etc., to ensure data quality);

[0072] S22, evaluate the weight and attribute volatility of environmental condition data to obtain key performance indicators;

[0073] S23, constructing several regression decision subtrees based on the random forest regression analysis algorithm, calculating the out-of-bag error, and constructing a prediction model;

[0074] S24, using the out-of-bag error (OOB error) and the running time of the prediction model as indicators to perform parameter optimization;

[0075] S25. Obtain several environmental condition attributes that have the greatest impact on water absorption performance.

[0076] In a further embodiment, evaluating the weight and attribute volatility of environmental condition data to obtain key performance indicators includes the following steps:

[0077] S221, obtaining the frequency of each environmental condition attribute by counting the number of times each environmental condition attribute appears in all environmental condition data, and comparing the number of times each environmental condition attribute appears with the total number of times all environmental condition attributes appear to obtain the weight of each environmental condition attribute;

[0078] S222, analyzing the volatility of each environmental condition attribute over time, that is, calculating the change in the number of occurrences of each environmental condition attribute at different time points. Attribute volatility can be determined by comparing the attribute occurrence frequency at consecutive time points (e.g., the previous day and the current day);

[0079] S223, obtaining a comprehensive score of each annular condition attribute by calculating the product of the volatility of the environmental condition attribute and the weight of the environmental condition attribute;

[0080] S224. Select several environmental condition attributes with the highest comprehensive scores as key performance indicators.

[0081] In a further embodiment, a plurality of regression decision subtrees are constructed based on a random forest regression analysis algorithm, and the out-of-bag error is calculated. The construction of the prediction model includes the following steps:

[0082] S231. Use the environmental condition data as a feature set and the water absorption performance data as a label set, and construct a data set through the feature set and the label set.

[0083] S232. Using the bagging idea and random subspace idea in random forest, several sample subsets and feature subspaces are randomly generated, and several regression decision subtrees are constructed.

[0084] It should be noted that the feature set contains the input variables, which are used by the prediction model to predict the target variable. In the matrix representation of data, the feature set corresponds to the column. The label set contains the output variable for each sample, that is, the value that the prediction model needs to predict. The label set is used to train the "answer" of the prediction model.

[0085] In random forests, sample subsets are obtained by sampling the original dataset with replacement and used to construct training data for each decision tree. Each sample subset includes samples randomly selected from the original dataset, which include both features and corresponding labels. The sample subset of each tree may overlap with the original dataset because it is sampled with replacement.

[0086] Feature subspace means that when constructing each split point of each decision tree, random forest does not use the entire feature set, but randomly selects a portion of the features from the feature set. That is, each tree may only consider a subset of the original feature set at each split point. This randomness helps the diversity of the model and reduces overfitting.

[0087] S233, train each regression decision subtree through the sample subset until the stopping condition is met (such as reaching the maximum depth, the number of samples in the node is less than the minimum number of split samples, etc.), and for each split node, search for the best split point through the feature subspace.

[0088] S234, calculating the out-of-bag error of each regression decision subtree, and taking the average of the out-of-bag errors of all regression decision subtrees as the overall out-of-bag error of the prediction model.

[0089] S235. All regression decision subtrees are combined to form a prediction model, which is obtained by averaging or voting the prediction results of all trees, and the generalization ability of the prediction model is evaluated through an independent test set (dividing the test set from the data set).

[0090] In a further embodiment, the parameter optimization using the out-of-bag error and the prediction model running time as indicators includes the following steps:

[0091] S241, select the number of features m when the initial split node is selected, and fix the value of m, obtain the change of the out-of-bag error and the prediction model training time with the number of regression decision subtrees, and find the number of regression decision subtrees that minimizes the out-of-bag error;

[0092] S242, select the number of regression decision subtrees obtained in S241, obtain the change of out-of-bag error and prediction model training time with the number of features when splitting the node, and find the number of features when splitting the node that minimizes the out-of-bag error;

[0093] S243, repeat S241-S242 until the out-of-bag error no longer decreases or reaches a preset threshold. This iterative method gradually adjusts the key parameters of the prediction model to minimize the OOB error while maintaining reasonable training efficiency. The meaning of OOB error is: each decision subtree is established by sampling with replacement (also known as bootstrap sampling) from the original training data set. This process will produce different subsets for training different decision subtrees. In the bootstrap sampling process, about 1 / 3 of the data will not be selected to build a specific decision subtree. This part of the unselected data is called "out-of-bag" data. Because these "out-of-bag" data do not participate in the training of the decision tree, they can be used to evaluate the performance of the tree, that is, to calculate the out-of-bag error.

[0094] In a further embodiment, obtaining several environmental condition attributes that have the greatest impact on water absorption performance comprises the following steps:

[0095] S251. Select several environmental condition attributes with the highest scores from the feature importance evaluation; wherein, in the random forest algorithm, the importance score of each feature can be directly obtained. These scores are usually based on the frequency of use of the features in the split nodes and their role in reducing the model error. Use the feature importance method built into the random forest algorithm, such as ranking based on average impurity reduction (Gini importance) or OOB error. Feature importance evaluation is a method used to measure the importance of each feature in the random forest algorithm. Feature importance helps understand which features contribute most to the predictive ability of the prediction model, so as to sort and select features.

[0096] S252. Use charts to visualize the importance scores of various environmental condition attributes.

[0097] S3. Retrain the prediction model based on the obtained key performance indicators and several environmental condition attributes.

[0098] In a further embodiment, retraining the prediction model based on the obtained key performance indicators and several environmental condition attributes includes the following steps:

[0099] S31, obtaining key performance indicators and several environmental condition attributes with the highest scores, and obtaining the final environmental condition selection attributes;

[0100] S32. Use the data of the final environmental conditions to select attributes to retrain the prediction model, adjust the prediction model parameters, and improve the accuracy of water absorption performance prediction.

[0101] S4. Use the retrained prediction model to predict the water absorption performance of the polymer water absorbent material, and optimize the formula or preparation process of the polymer water absorbent material based on the prediction results.

[0102] In a further embodiment, according to the prediction results, optimizing the formula or preparation process of the polymer water-absorbing material comprises the following steps:

[0103] S41, constructing a parameter optimization model according to the prediction results of the prediction model and a combination of several randomly selected recipes or preparation process parameters;

[0104] S42, performing iterative optimization based on the acquisition function, and updating the parameter optimization model;

[0105] S43. Select the formula or process parameters with the best water absorption performance predicted by the parameter optimization model as the final optimization result.

[0106] In a further embodiment, constructing a parameter optimization model according to the prediction results of the prediction model and a combination of several randomly selected recipes or preparation process parameters includes the following steps:

[0107] S411. Construct an objective function, and the objective function is:

[0108]

[0109] In the formula, f ( x ) represents the objective function value, x A vector representing all recipe or process parameters;

[0110] SR ( x ) means according to the parameter x Predicted water absorption rate, MWC ( x ) means according to the parameter x Predicted maximum water absorption, WR ( x ) means according to the parameter x predicted water retention capacity;

[0111] w 1 , w 2 , w 3 Both represent weight coefficients, which are used to balance the importance of different performance indicators;

[0112] S412, randomly selecting a combination of several formulas or process parameters, and using the prediction model to predict, to obtain data on water absorption performance, and constructing an initial data set based on the randomly selected combinations of several formulas or process parameters and the obtained data on water absorption performance;

[0113] S413, using the initial data set to establish a parameter optimization model, and using the parameter optimization model to predict the water absorption performance of the polymer water-absorbing material with unknown formula or process parameter combination. The parameter optimization model is based on a Gaussian regression model, which is a non-parametric probability model used to predict unknown data points based on a given data set. In the present invention, the parameter optimization model is used to predict the water absorption performance under different formulas or process parameter combinations.

[0114] When training a parameter optimization model, a kernel function is selected. The kernel function defines the similarity between data points. Commonly used kernel functions include radial basis functions, etc.

[0115] Use the initial dataset to train the parameter optimization model, that is, learn the parameters (hyperparameters) of the kernel function and the noise parameters. Optimize the hyperparameters of the kernel function by maximizing the marginal likelihood or using cross-validation. At the same time, evaluate the parameter optimization model with an additional validation dataset.

[0116] In a further embodiment, performing iterative optimization based on the acquisition function and updating the parameter optimization model comprises the following steps:

[0117] S421. Select an acquisition function, such as expected improvement (EI) or upper confidence bound (UCB), and determine the next formulation or process parameter combination with the greatest potential for improving water absorption performance through the acquisition function.

[0118] S422. Use the acquisition function to select the next experimental plan, and add the new water absorption performance data to the initial data set to form an experimental data set.

[0119] It should be noted that the experimental scheme refers to a specific set of input parameters, which are the parameters of the formulation and preparation process of the high-analysis water-absorbing material. Each experimental scheme represents a possible formulation or process setting that will affect the performance of the polymer water-absorbing material. Use the acquisition function to select an experimental scheme with the greatest potential for improving performance based on the current parameter optimization model. Implement this scheme, that is, prepare the material according to these parameters and test its water absorption performance. Add the obtained performance data (such as water absorption rate, maximum water absorption, water retention capacity, etc.) to the data set together with the experimental scheme.

[0120] S423: Use the experimental data set to update the parameter optimization model, and repeat the updating of the parameter optimization model until a preset number of iterations is reached.

[0121] In summary, the data collection and analysis method for detecting polymer water-absorbing materials provided by the present invention builds a prediction model based on a random forest regression algorithm, which is used to predict the water absorption performance of polymer water-absorbing materials, and optimizes parameters by using the out-of-bag error and the running time of the prediction model as indicators to improve the performance of the prediction model and shorten the running time; by evaluating the weight and volatility of environmental conditions, identifying and quantifying the contribution of each environmental condition to the performance, helping to focus on the most critical factors, and determining the environmental factors that have the greatest impact on water absorption performance, and retraining the prediction model by focusing on the data of key performance indicators and environmental condition attributes, the prediction accuracy and generalization ability of the model are improved. The present invention constructs a parameter optimization model based on the prediction results of the prediction model and a combination of several randomly selected formulas or preparation process parameters, systematically optimizes the material formula or preparation process, finds the optimal formula or process parameters of high-analysis water-absorbing materials, maximizes performance, and reduces costs.

[0122] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A data collection and analysis method for detecting polymer water-absorbing materials, It is characterized in that The data collection and analysis method for detecting polymer water-absorbing materials comprises the following steps: S1. Collect water absorption performance data of polymer water-absorbing materials under different environmental conditions; S2. Analyze the water absorption performance data, obtain key performance indicators, and build a prediction model to screen out several environmental condition attributes that have the greatest impact on water absorption performance; S3. Retraining the prediction model based on the obtained key performance indicators and several environmental condition attributes; S4. Predicting the water absorption performance of the polymer water absorbent material using the retrained prediction model, and optimizing the formula or preparation process of the polymer water absorbent material according to the prediction result; The steps of analyzing the water absorption performance data, obtaining key performance indicators, and constructing a prediction model to screen out several environmental condition attributes that have the greatest impact on the water absorption performance include the following steps: S21, quantifying and preprocessing the non-numerical features in the water absorption performance data and the environmental condition data; S22, evaluate the weight and attribute volatility of environmental condition data to obtain key performance indicators; S23, constructing several regression decision subtrees based on the random forest regression analysis algorithm, calculating the out-of-bag error, and constructing a prediction model; S24, using the out-of-bag error and the running time of the prediction model as indicators to perform parameter optimization; S25. Obtaining several environmental condition attributes that have the greatest impact on water absorption performance; The step of evaluating the weight and attribute volatility of environmental condition data to obtain key performance indicators includes the following steps: S221, obtaining the frequency of each environmental condition attribute by counting the number of times each environmental condition attribute appears in all environmental condition data, and comparing the number of times each environmental condition attribute appears with the total number of times all environmental condition attributes appear to obtain the weight of each environmental condition attribute; S222. Analyze the volatility of each environmental condition attribute over time; S223, obtaining a comprehensive score of each annular condition attribute by calculating the product of the volatility of the environmental condition attribute and the weight of the environmental condition attribute; S224. Select several environmental condition attributes with the highest comprehensive scores as key performance indicators; The method of constructing a plurality of regression decision subtrees based on the random forest regression analysis algorithm and calculating the out-of-bag error to construct a prediction model includes the following steps: S231, using the environmental condition data as a feature set and the water absorption performance data as a label set, and forming a data set through the feature set and the label set; S232. Using the bagging idea and random subspace idea in random forest, several sample subsets and feature subspaces are randomly generated, and several regression decision subtrees are constructed; S233, training each regression decision subtree through the sample subset until the stopping condition is met, and for each split node, searching for the best split point through the feature subspace; S234, calculating the out-of-bag error of each regression decision subtree, and taking the average of the out-of-bag errors of all regression decision subtrees as the overall out-of-bag error of the prediction model; S235, combining all regression decision subtrees to form a prediction model, and evaluating the generalization ability of the prediction model through an independent test set; The parameter optimization using the out-of-bag error and the prediction model running time as indicators includes the following steps: S241, select the number of features m when the initial split node is selected, and fix the value of m, obtain the change of the out-of-bag error and the prediction model training time with the number of regression decision subtrees, and find the number of regression decision subtrees that minimizes the out-of-bag error; S242, select the number of regression decision subtrees obtained in S241, obtain the change of out-of-bag error and prediction model training time with the number of features when splitting the node, and find the number of features when splitting the node that minimizes the out-of-bag error; S243, repeat S241-S242 until the out-of-bag error no longer decreases or reaches a preset threshold; The method of obtaining the several environmental condition attributes that have the greatest impact on the water absorption performance comprises the following steps: S251. Select several environmental condition attributes with the highest scores from the feature importance evaluation; S252. Use charts to visualize the importance scores of various environmental condition attributes.

2. A data collection and analysis method for detecting polymer water-absorbing materials according to claim 1, It is characterized in that The retraining of the prediction model based on the obtained key performance indicators and several environmental condition attributes includes the following steps: S31, obtaining key performance indicators and several environmental condition attributes with the highest scores, and obtaining the final environmental condition selection attributes; S32. Use the data of the final environmental condition selection attributes to retrain the prediction model.

3. A data collection and analysis method for detecting polymer water-absorbing materials according to claim 1, It is characterized in that The method of optimizing the formula or preparation process of the polymer water-absorbing material according to the prediction results comprises the following steps: S41, constructing a parameter optimization model according to the prediction results of the prediction model and a combination of several randomly selected recipes or preparation process parameters; S42, performing iterative optimization based on the acquisition function, and updating the parameter optimization model; S43. Select the formula or process parameters with the best water absorption performance predicted by the parameter optimization model as the final optimization result.

4. A data collection and analysis method for detecting polymer water-absorbing materials according to claim 3, It is characterized in that The method of constructing a parameter optimization model based on the prediction results of the prediction model and a combination of several randomly selected recipes or preparation process parameters comprises the following steps: S411. Construct an objective function, and the objective function is: In the formula, f ( x ) represents the objective function value, x A vector representing all recipe or process parameters; SR ( x ) means according to the parameter x Predicted water absorption rate, MWC ( x ) means according to the parameter x Predicted maximum water absorption, WR ( x ) means according to the parameter x predicted water retention capacity; w 1 , w 2 , w 3 All represent weight coefficients; S412, randomly selecting a combination of several formulas or process parameters, and using the prediction model to predict, to obtain data on water absorption performance, and constructing an initial data set based on the randomly selected combinations of several formulas or process parameters and the obtained data on water absorption performance; S413. Establish a parameter optimization model using the initial data set, and predict the water absorption performance of a polymer water-absorbing material with an unknown formula or process parameter combination through the parameter optimization model.

5. A data collection and analysis method for detecting polymer water-absorbing materials according to claim 4, It is characterized in that The iterative optimization based on the acquisition function and updating the parameter optimization model include the following steps: S421, selecting an acquisition function, and determining the next formula or process parameter combination for improving water absorption performance through the acquisition function; S422, using the acquisition function to select the next experimental scheme, and adding the new water absorption performance data to the initial data set to form an experimental data set; S423: Use the experimental data set to update the parameter optimization model, and repeat updating the parameter optimization model until a preset number of iterations is reached.

Images