RF-meta model-based agricultural mulch film degradation period prediction method and device

The RF-Meta model established through Meta analysis and random forest algorithm solves the accuracy and universality of BDM degradation cycle prediction, realizes consideration of multi-factor interaction, provides globally applicable degradation cycle prediction, and promotes the environment Protection and resource conservation.

WO2024234733A1PCT designated stage expired Publication Date: 2024-11-21JIANGSU ACAD OF AGRI SCI

Patent Information

Application Number
PCT/CN2024/074429
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2024-01-25
Filing Date
2024-01-29
Publication Date
2024-11-21

AI Technical Summary

Technical Problem

Existing research is difficult to accurately predict the degradation cycle of biodegradable plastic mulch (BDM), and it is impossible to comprehensively analyze the nonlinear interaction of materials, soil, climate and human factors, resulting in difficulty in predicting the degradation process.

Method used

The Meta analysis method combined with the random forest algorithm is used to establish an RF-Meta model. By collecting and verifying the original data, filling in missing values, dividing the data set, training and optimizing the model, the prediction curve of the degradation cycle is obtained.

Benefits of technology

Accurate estimation of BDM degradation cycles is achieved, the versatility and applicability of model results are improved, complex interactions of multiple factors can be considered, and mulch degradation predictions are provided globally, promoting environmental protection and resource conservation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN2024074429_21112024_PF_FP_ABST
    Figure CN2024074429_21112024_PF_FP_ABST
Patent Text Reader

Abstract

Disclosed in the invention is an RF-Meta model-based agricultural mulch film degradation period prediction method, comprising the following steps: by means of a Meta analysis method, collecting a biodegradable mulch film degradation original data set; by means of Meta analysis, sorting the influence QM of different factors in the original data set on the degradation degree; screening QMs meeting required factors for keeping in the database, and using a random forest algorithm, filling missing values in the original data set, so as to obtain a complete database; dividing the complete database into a modeling set and a verification set, and establishing an initial RF-Meta model; training the initial RF-Meta model and optimizing hyper-parameters, so as to obtain an optimized RF-Meta model; importing initial experimental physicochemical data of the degradation of a target biodegradable mulch film into the optimized RF-Meta model, so as to obtain a prediction curve of the full degradation period. The model provided in the invention can achieve, for the first time, specific degradation degree numerical prediction for the degradation process of BDM, and has wide applicability.
Need to check novelty before this filing date? Find Prior Art

Description

A method and device for predicting agricultural mulch degradation cycle based on RF-Meta model Technical Field

[0001] The present invention relates to the field of agricultural informatization, and in particular to a method and device for predicting the soil-buried degradation period of biodegradable mulch films by combining a random forest algorithm with a meta-analysis method. Background Art

[0002] Mulch film is a crucial agricultural resource. It effectively suppresses weed growth, increases crop yields, maintains soil moisture, reduces evaporation, and conserves water resources. Furthermore, it raises soil temperature, promotes crop growth, reduces soil erosion, and maintains soil fertility, thereby extending the crop life cycle, reducing the risk of pests and diseases, reducing the use of pesticides and fertilizers, and improving food quality. China uses a staggering 137.9 million tons of mulch film annually. However, due to ultraviolet radiation and external forces, mulch film is prone to fragmentation, forming plastic fragments that cause serious plastic pollution in farmland and pose a potential threat to ecosystems and human health. The impact of plastic pollution is currently receiving increasing attention from the scientific community and the public. Biodegradable plastics are plastics that can be degraded and utilized by microorganisms, ultimately converting them into microbial biomass, carbon dioxide, and water. Therefore, replacing polyethylene (PE) mulch with biodegradable plastic mulch film (BDM) is a key strategy for preventing plastic pollution in agriculture. However, the decomposition cycle of BDM in the environment is still unclear. This is mainly because the degradation of BDM is a complex process and is closely related to the material composition and natural conditions. The degradation effects of BDM with different formulations in different regions are often very different.

[0003] At present, the research on the degradation cycle of specific BDM mulch film products is mainly based on the experience of inferring its degradation cycle after 1 to 2 years of use in local farmland. There have been relevant reports on the effects of material factors, soil factors, regional factors, etc. on the degradation of BDM, but the relevant research mainly relies on experimental data and can only give a judgment on the general trend of degradation. The research methods are relatively simple and limited by the number of control groups in the experiment. It is difficult to measure the influence of relevant factors on degradation, and it is also impossible to make relevant predictions on the degradation cycle. Existing studies have established a simple fitting model for the ultraviolet radiation (UV) on the degradation cycle of mulch film to evaluate the photodegradation cycle of mulch film. The shortcomings of the above research are: (1) UV is not the only relevant influencing variable of BDM; (2) The degradation process of BDM is complex, and material factors, soil factors, climate factors, and human factors often show nonlinear interactions. Analyzing a small number of factors cannot fully analyze this interactive characteristic and cannot achieve a specific prediction of the degradation process; (3) Current research cannot solve the regional characteristics of mulch film degradation. Therefore, it is urgent to develop a nonlinear BDM degradation prediction model with richer variables and wider applicability in the mulch film use area.

[0004] Summary of the Invention

[0005] Purpose of the invention: The purpose of the present invention is to provide an RF-Meta model that combines the Meta analysis method and the random forest algorithm to predict the degradation period of agricultural mulch film, achieve accurate estimation of the degradation period, and improve the versatility of the model results.

[0006] Technical solution: A method for predicting the degradation cycle of agricultural mulch films based on the RF-Meta model, including the following steps:

[0007] (1) The original data set of biodegradable mulch degradation was collected through meta-analysis, an original database was established based on the original data set, and the reliability of the original data was verified through meta-analysis.

[0008] (2) Meta-analysis of the effects of different factors on the degradation degree in the original data set M Sort by

[0009] (3) Based on the results of Meta-analysis, Q M Factors with values ​​greater than 1 were retained in the original database, and the random forest algorithm was used to fill in the missing values ​​in the original database to obtain a complete database;

[0010] (4) Divide the complete database into a modeling set and a validation set, and establish the initial RF-Meta model;

[0011] (5) Train the initial RF-Meta model and optimize the hyperparameters to obtain the optimized RF-Meta model;

[0012] (6) The initial experimental physicochemical data of the target biodegradable mulch film degradation were imported into the optimized RF-Meta model to obtain the prediction curve of the entire degradation cycle.

[0013] Furthermore, the specific steps of constructing the original database in step (1) are as follows:

[0014] (1.1) Develop a search strategy and inclusion and exclusion criteria according to the PICOS guidelines, including:

[0015] P: The research object is determined to be biodegradable mulch in agricultural production scenarios;

[0016] I: Determine intervention measures and select the range of basic material components for degradable mulch; and screen different degradation methods, degradation environments, and pre-degradation treatments;

[0017] C: Determine the control measure as the film properties before and after degradation;

[0018] O: Determine the result orientation as the film degradation rate parameter;

[0019] S: The research design was limited to laboratory experiments or field experiments, and reviews and opinion articles were excluded;

[0020] (1.2) Select an article database, establish search terms according to the criteria in (1.1), and manually screen articles that meet the research criteria and can be replicated;

[0021] (1.3) The feature data set G1 to be collected in each article is composed as follows: G1 = {A1, A2..., A a ; C1, C2, ..., C e ;SP1,SP2,...,SP p ;SB1,SB2...,SB b ;M1,M2...,M m ;E1,E2...,E e}; Among them, A a Represents item a in the basic information of the article to be collected, C e Represents the eth item in the climate conditions to be collected, SP p represents the pth item in the soil physical condition to be collected, SB b Represents item b of the biological and chemical conditions of the soil to be collected, M m Represents the mth item in the characteristics of the ground film material to be collected, E e represents the e-th item in the experimental conditions to be collected;

[0022] (1.4) The composition of the degradation result data set G2 to be collected in each article is as follows: G2 = {E m ,E v ,C m ,C v}; where E m represents the mean of the experimental group results, E v represents the variance of the experimental group results, C m represents the mean of the control group, C v represents the variance of the control group results; the formulas for the mean and standard deviation are as follows:

[0023] in, is the mean of the degradation rate, s is the standard deviation of the degradation rate; x i is the degradation rate of group i, where i belongs to 1.2.3...n.

[0024] Furthermore, the verification of the reliability of the original data by Meta-analysis includes:

[0025] (1.5) By observing the relationship between the degradation rate and its variance, we can test whether the collected articles have publication bias, that is, the data source has obvious tendency; the degradation rate of the biodegradable mulch obtained from each set of experimental data is ranked according to the degradation rate, and then the position number u of each data is obtained. i ; Then sort each group of data according to the size of the experimental variance, and obtain the position number v after sorting j Then calculate the correlation coefficient D between the degradation rate and its experimental variance and the probability P of the hypothesis represented by D: P = 2 × (1 - cdf (D))

[0026] Among them, n is the total amount of data in the database, C n represents the number of concordant pairs in pairwise comparisons, D n is the number of discordant pairs in the pairwise comparison; if u of two variables i and v j Increase or decrease at the same time, they are a concordant pair, the u of the two variables i and v j If one variable increases while the other decreases, they are a discordant pair. cdf(D) is the cumulative distribution function of D under the standard normal distribution. The presence of publication bias can be determined by referring to the D value and P value. If the absolute value of the D value is greater than 1.96 and the p value is less than 0.05, the data is considered to have a clear bias and the data source with a clear bias needs to be excluded. If there is no clear bias, proceed to the next step.

[0027] (1.6) When publication bias is found, first use R language to draw a funnel plot with degradation rate as x-axis data and experimental variance as y-axis data. The position vector of each data in the plot is d i =(x i ,s i );by For the axis, if the numeric vector d t =(x t ,s t ) There is obviously no data near the corresponding point on the other side of the axis, resulting in an asymmetric funnel plot. Then further determine whether the data vector corresponds to a perfect experimental design, the experimental environment meets the requirements, and the IF of the journal included in the paper is greater than 6. If it meets the requirements, add the vector to the database. To eliminate the bias; if the data vector corresponds to an experiment with a special experimental environment or low reliability, delete the data from the database; after screening the data that cause the funnel plot to be asymmetric, repeat (1.6) and recheck the publication bias;

[0028] (1.7) Evaluate whether there are general patterns in the data in the database; the data significance level is calculated as:

[0029] Data significance tendency level F n When F is greater than zero, the data in the database are considered to have a significant general law. n <0, then you need to go back to step (1.1) and add intervention measures to the selection of articles to ensure that the data will have significant patterns.

[0030] Furthermore, step (2) is specifically as follows:

[0031] (2.1) Scan the degradation rate data and variance data to check for missing or null values. Once a null value is found, record its location information, including the row m and column index n where the null value is located, and fill it in using the nearest neighbor linear interpolation method.

[0032] (2.2) A multivariate random effects model for meta-analysis was established for the database. The basic random effects of the model were generated by the different literature sources from which the data were extracted. The formula of the multivariate random effects model was: yi = Xiβ + ui + ∈i

[0033] Where: yi is the effect size of the ith study, Xi is the covariate / predictor variable associated with the ith study, β is the regression coefficient of the model, ui is the random effect of the ith study, representing the heterogeneity between studies, ∈i is the error term of the model;

[0034] (2.3) The weight of each feature’s influence on the result is calculated by the following formula:

[0035] Qm is the weight used to measure the impact of each feature on the result. j represents the Qm calculated for the jth feature, and k represents the number of data in the database that contains this feature. is the weight of the jth feature factor, x ·j is the degradation rate of the jth characteristic factor.

[0036] Furthermore, it is characterized in that the step (3) is specifically:

[0037] 1) Initial state: In the original database established by the Meta-analysis method, some data values ​​are empty, and the jth feature data of the i-th data in the database is recorded as A ij ;

[0038] 2) Random Forest Architecture: The data is organized according to the different types of feature data to be collected in step (1.3) (A, C, SP, SB, M, E). Each feature category has different feature data. For each feature category, an independent decision tree model is established for each feature, and s in different feature categories are randomly extracted. features, establish d independent decision tree models to simulate the mutual influence of features under different feature classifications, and finally establish a total of d+6 decision tree models to form an ensemble of random forest models;

[0039] 3) Extracting data relationship features: Each decision tree will judge and classify each feature j, and learn the data relationship under the specific feature classification from the data; the data relationship features include the value range of the feature, the relationship between the features, and the data pattern;

[0040] 4) Progressive filling: traverse all the vacant values ​​in the database and record the number of all vacant values ​​under different feature j columns as θ j , establish the matrix [θ1,θ2,…θ j ], and check whether there is a value less than or equal to 0. If there is, remove the data; then use the bubble sort method to sort the θ values ​​in the matrix from small to large; then fill all the data under the j feature in order from small to large according to the sorted results; finally traverse the matrix [θ1,θ2,…θ j ], traverse the data in the database again to ensure that there are no missing values;

[0041] 5) Regression prediction: Based on the experimental group data with complete eigenvalues ​​in the original database, the random forest algorithm is used. Using the d+6 decision trees established in 2), the values ​​judged by the d+6 decision trees are calculated respectively, and the average is taken to fill the missing values ​​in sequence.

[0042] Furthermore, in step (4), the complete data set is divided into a modeling set and a validation set by random sampling.

[0043] Furthermore, the RF-Meta model was optimized through five-fold cross validation.

[0044] Furthermore, the hyperparameter optimization uses GridSearchCV to perform grid multi-core accelerated search to find the optimal hyperparameters.

[0045] A device for predicting the degradation cycle of agricultural mulch film based on the RF-Meta model, comprising a processor and an executable program stored in a memory and runnable on the processor, characterized in that when the processor executes the executable program, a method for predicting the degradation cycle of agricultural mulch film based on the RF-Meta model is implemented.

[0046] A storage medium containing a computer executable program, characterized in that when the computer executable program is executed by a computer processor, it is used to execute a method for predicting the degradation cycle of agricultural film based on the RF-Meta model.

[0047] Beneficial Effects: 1. Accurate Estimation of Degradation Cycles: The degradation of mulch films requires consideration of numerous parameters, such as temperature, humidity, soil type, and microbial activity. The model examines the complex interactions among these parameters, resulting in a precise estimate of the degradation cycle. 2. Improved Universality of Model Results: By collecting experimental data from major mulch film use sites around the world, the developed model is applicable to global BDM degradation scenarios. 3. Accelerated Degradation: The application of the predictive model enables farmers to better understand the relationship between key parameters such as soil moisture and total soil nitrogen and the degradation rate of mulch films. This allows them to take measures to adjust these parameters to accelerate mulch film degradation, thereby promoting environmental protection and reducing plastic residue on the land. 4. Scientifically Selected Mulch Films: The predictive model also provides farmers with a scientific basis for selecting the type of mulch film best suited to local soil and climatic conditions. This helps improve the efficiency of mulch film use, reduces waste and resource consumption, and ensures healthy crop growth. 5. Reduced Labor Costs: With the known degradation cycle of mulch films, farmers no longer need to employ large numbers of people to clean up mulch film debris from their fields. Farmers can plan their farmland more flexibly, better plan mulch replacement and crop rotation, and achieve sustainable ecological agricultural production. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] FIG1 is a flow chart of model establishment of the present invention;

[0049] Figure 2 shows the prediction accuracy of the independent validation set evaluated by the coefficient of determination;

[0050] Figure 3 is a funnel plot of the Egger regression test results. DETAILED DESCRIPTION

[0051] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0052] As shown in Figure 1, the model building process of the present invention includes the following steps:

[0053] (1) Collect the original data set of biodegradable mulch degradation through meta-analysis method;

[0054] (2) Use the random forest model to fill in the missing values ​​in the original dataset to obtain a complete database;

[0055] (3) Divide the complete database into a modeling set and a validation set, and establish the initial RF-Meta model.

[0056] (4) Train the initial RF-Meta model and optimize the hyperparameters to obtain the optimized RF-Meta model;

[0057] (5) Optimize the hyperparameters of the optimized random forest prediction model to obtain the optimal RF-Meta model;

[0058] (6) Verify the accuracy of the random forest prediction model.

[0059] Example:

[0060] 1. Obtaining the BDM degradation dataset: The raw data for model building were collected and integrated using the PRISMA workflow of the meta-analysis method. Peer-reviewed publications that investigated the effects of climate, materials, and soil environment on the degradation of biodegradable films over a 10-year period and the China National Knowledge Infrastructure database were searched using the Web of Science website. To maintain quality control, criteria for data inclusion were established: (1) Experimental conditions should be similar to the natural degradation environment, and no additional strains should be added to the soil environment for cultivation. The degradation temperature of the soil (except compost) should not exceed 35°C. (2) The main components of the BDM materials used in the experiment need to be reported, and the proportion of PLA and PBAT should be greater than 30%. (3) The experimental data are replicated, and information on the degraded soil in the experiment can be obtained.

[0061] Variables collected from these studies included: article title; publication year; study location; meteorological variables, including cumulative rainfall, average temperature, and solar radiation during the degradation period; soil properties, including initial soil moisture content, total soil nitrogen content, soil organic matter, pH, clay content, silt content, sand content, bulk density, and soil coarse particle content; material variables, including PLA content, PBAT content, material thickness, and material surface area; degradation time; burial depth; and soil ecological variables, including the Chao1 index and the Shannon index. Furthermore, we considered whether the BDM had been exposed to outdoor conditions for more than seven days prior to interpretation and whether it had been plowed during degradation.

[0062] We supplemented some missing values ​​with known databases, including the China Global Land Reanalysis 40-Year Product (CRA / Land) - Monthly Product (Land Product) and the SoilGrids database. The BDM degradation rate is calculated as follows:

[0063] This formula represents the degradation rate obtained by calculating the mass drop before and after degradation. b and W t Represents the weight of the specimen before burial in soil and after extraction from the soil at the end of degradation.

[0064] This equation gives the percentage of current degradation level by dividing the amount of CO2 released by the amount of CO2 that would theoretically be released if the BDM were fully biodegraded. (CO2) t The value of represents all CO2 emissions generated during the biodegradation of BDM, while (CO2) b Corresponding to the blank experiment without BDM addition, it represents the CO2 emission in the soil environment without BDM. t is the total dry weight of biodegradable plastic material, C t is the relative weight of the total carbon in the sample, 44 and 12 are the molecular mass of CO2 and the atomic mass of carbon.

[0065] The funnel plot of the model's Egger regression test results is shown in Figure 3. The plot is symmetrical on both sides, with the data points distributed at the bottom of the funnel, rather than exhibiting significant asymmetry or deviation from a straight line. This indicates that the data are not significantly biased, and the results are highly reliable. Using the Rosenthal method, the failure safety factor was calculated to be 87,778, indicating that approximately 87,778 nonsignificant studies would be required to offset the observed statistical significance, indicating that the data are clearly significant.

[0066] 2. A multivariate random effects model (meta-regression) was fitted using the rma.mv function in the metafor package in R. The collected factors were ranked according to the heterogeneity QM caused by each factor. The factors with QM > 1 were retained as follows: cumulative rainfall, average temperature, initial soil moisture content, soil total nitrogen content, soil organic matter, pH, clay content, mud content, sand content, bulk density, soil coarse particle content, PLA content, PBAT content, burial depth, Chao1 index, Shannon index, and degradation time.

[0067] 3. Filling the database data: Given the complexity of our data sources, resulting in 8% of the data containing null values, before officially starting modeling, we used the same method used for the final modeling to fill in the small amount of null values ​​in the database, without changing the final result. We adopted an intelligent, step-by-step approach. The core idea of ​​this method is to traverse all features and fill in the missing data starting with the feature with the least, thus ensuring minimal changes to the original data features. The filling process is as follows:

[0068] (1) Initial state: Our database contains multiple features, some of which may have missing values. Before the filling process begins, we first select the feature with the least missing values ​​and fill the missing values ​​of other features with 0.

[0069] (2) Regression prediction: We use known data to perform random forest regression prediction, and estimate the possible value of the missing feature through the data relationship of other complete data. This estimated value is consistent with the characteristics of the overall data. This predicted value will be put back into the original feature matrix to gradually improve the data.

[0070] (3) Progressive filling: After we successfully fill in a feature, we continue to select the next feature with the least number of missing values ​​and repeat the above steps until all features are traversed.

[0071] Finally, we filled all the empty values ​​in the database with high-confidence values ​​based on the existing information.

[0072] 4. Constructing a modeling set and an independent validation set: The entire dataset collected above was used as the sample set, with the degradation rate as the fitting target and the remaining data as feature data. The sample set was divided into a modeling set and an independent validation set. Through random sampling, 70% of the data in the dataset was assigned to the modeling set, and the remaining 30% was assigned to the independent validation set. In this example, the total dataset contained 732 data samples.

[0073] 5. Training the initial random forest prediction model: Using the BDM degradation rate data and many screened variables in the modeling set as training data, the initial random forest prediction model is trained, and the parameters of the initial random forest prediction model are optimized through five-fold cross validation to obtain the random forest prediction model.

[0074] The random forest prediction model provided by the present invention is a data mining algorithm that uses bootstrap sampling (with replacement) to generate n (n is usually set to 500) data sets of the same size as the training sample set, and is used to construct n decision trees. The environmental variables are randomly divided into multiple environmental variable subsets. In each decision tree, the nodes are branched by randomly dividing the multiple environmental variable subsets. The final prediction result of the model is the average of the prediction results of all decision trees.

[0075] Hyperparameter Optimization: Optimize the model's hyperparameters to achieve the best fit. We used GridSearchCV to perform a multi-core accelerated search to find the optimal hyperparameters. The final hyperparameter values ​​obtained were: n_estimators = 186, random_state = 42, and max_depth = 23.

[0076] 6. Verify the random forest prediction model: The established random forest prediction model is used to predict the independent validation set, and the actual degradation rate in the independent validation set is compared to evaluate the prediction accuracy of the random forest conversion function. In this embodiment, the prediction accuracy of the independent validation set is evaluated by the coefficient of determination (R2). The results of the evaluation are shown in Figure 2. As can be seen from Figure 2, the R2 of the independent validation set is 0.965, which has a very good prediction effect. We also verified the root mean square error (RMSE) of the data, which is 6.332, indicating that the mean prediction error of our model is 6.3%. In view of the complexity of the agricultural scenario, it can be considered that the model we established can well predict the BDM degradation rate and has practical prospects.

[0077] Meta-analysis is a statistical method used to synthesize the results of multiple independent studies to obtain stronger, more comprehensive conclusions. Meta-analysis can identify patterns, effect sizes, and credibility, helping decision makers reach more reliable conclusions. It can reveal general patterns across studies while accounting for the weights and biases of individual studies, enhancing the comprehensiveness and credibility of the data.

[0078] Random forests are a tree-based machine learning method that effectively handles nonlinear relationships and exhibits robustness for most regression problems. Furthermore, regression forests can effectively avoid overfitting and are suitable for building models with relatively small amounts of data. Therefore, combining meta-analysis with random forest models is promising for establishing a predictive model for BDM degradation.

[0079] The model proposed in the present invention is the first to provide a specific numerical prediction of the degree of degradation of BDM during its degradation process, and has wide applicability and can be used for BDM prediction in different soil, climate, and material environments. Furthermore, by screening and processing the original information through meta-analysis, it can be ensured that the data set will not mislead the model. In a new scenario, we only need to provide the initial physical and chemical information of the degradation scenario to predict its future degradation capacity. Compared with the conventional testing method of obtaining the BDM degradation characteristics of local farmland through field experiments, we do not need to spend a lot of manpower and material resources, and do not need a long experimental period, providing a new approach for the prediction of the BDM degradation cycle.

Claims

1. A method for predicting agricultural mulch degradation cycle based on RF-Meta model. It is characterized in that The following steps are involved: (1) The original data set of biodegradable mulch degradation was collected through meta-analysis, an original database was established based on the original data set, and the reliability of the original data was verified through meta-analysis. (2) Meta-analysis of the effects of different factors on the degradation degree in the original data set M Sort by (3) Based on the results of the meta-analysis, select M The factors with set value ranges were kept in the original database, and the random forest algorithm was used to fill the missing values ​​in the original database to obtain a complete database; (4) Divide the complete database into a modeling set and a validation set to establish the initial RF-Meta model; (5) Train the initial RF-Meta model and optimize the hyperparameters to obtain the optimized RF-Meta model; (6) The initial experimental physicochemical data of the target biodegradable mulch film degradation were imported into the optimized RF-Meta model to obtain the prediction curve of the entire degradation cycle.

2. According to claim 1, a method for predicting agricultural mulch film degradation cycle based on RF-Meta model, It is characterized in that The specific steps of constructing the original database in step (1) are as follows: (1.1) Develop a search strategy and inclusion and exclusion criteria according to the PICOS guidelines, including: P: The research object is determined to be biodegradable mulch in agricultural production scenarios; I: Determine intervention measures and select the range of basic material components of degradable mulch; and screen different degradation methods, degradation environments, and pre-degradation treatments; C: Determine the control measure as the characteristics of the mulch before and after degradation; O: Determine the result orientation as the film degradation rate parameter; S: The research design was limited to laboratory experiments or field experiments, and reviews and opinion articles were excluded; (1.2) Select an article database, establish search terms according to the criteria in (1.1), and manually screen articles that meet the research criteria and can be replicated; (1.3) The feature data set G to be collected in each article 1 The composition of G is as follows: 1 ={A 1 ,A 2 ...,A a ; C 1 ,C 2 ,...,C e ;SP 1 ,SP 2 ,...,SP p SB 1 ,SB 2 ...,SB b ;M 1 ,M 2 ...,M m ; E 1 ,E 2 ...,E e }; Among them, A a represents the ath item in the basic information of the article to be collected, C e Represents the eth item in the climate conditions to be collected, SP p represents the pth item in the soil physical condition to be collected, SB b Represents item b of the biological and chemical conditions of the soil to be collected, M m represents the mth item in the characteristics of the ground film material to be collected, E e represents the e-th item in the experimental conditions to be collected; (1.4) The degradation result data set G to be collected in each article 2 The composition of G is as follows: 2 ={E m ,E v ,C m ,C v }; where E m represents the mean of the experimental group results, E v represents the variance of the experimental group results, C m represents the mean of the control group, C v represents the variance of the control group results; the formulas for the mean and standard deviation are as follows: in, is the mean of the degradation rate, s is the standard deviation of the degradation rate; x i is the degradation rate of the i-th group, i belongs to 1.2.3...n.

3. The method for predicting agricultural mulch film degradation cycle based on RF-Meta model according to claim 2, It is characterized in that The verification of the reliability of the original data by Meta-analysis includes: (1.5) By observing the relationship between the degradation rate and its variance, we can test whether the collected articles have publication bias, that is, the data source has an obvious tendency; the degradation rate of the biodegradable mulch obtained from each group of experimental data is ranked according to the degradation rate, and then the position number u of each data is obtained. i ; Then sort each group of data according to the size of the experimental variance, and obtain the position number v after sorting j Then calculate the correlation coefficient D between the degradation rate and its experimental variance and the probability P of the hypothesis represented by D: P = 2 × (1-cdf (D)) Where n is the total amount of data in the database, C n represents the number of concordant pairs in pairwise comparisons, D n is the number of discordant pairs in the pairwise comparison; if the u of the two variables i and v j Increase or decrease at the same time, they are a concordant pair, and the u of the two variables i and v j If one variable increases while the other decreases, they are a discordant pair; cdf(D) is the cumulative distribution function value of D value under standard normal distribution; by referring to D value and P value, we can know whether there is publication bias. If the absolute value of D value is greater than 1.96 and the p value is less than 0.05, it is considered that the data has obvious bias and it is necessary to exclude the data source with obvious bias; if there is no obvious bias, proceed to the next step; (1.6) When publication bias is found, first use R language to draw a funnel plot with degradation rate as x-axis data and experimental variance as y-axis data. The position vector of each data in the plot is d i =(x i ,s i );by For the axis, if the numeric vector d t =(x t ,s t ) There is obviously no data near the corresponding point on the other side of the axis, resulting in an asymmetric funnel plot. Then further determine whether the data vector corresponds to a perfect experimental design, the experimental environment meets the requirements, and the IF of the journal included in the paper is greater than 6. If it meets the requirements, add the vector to the database To eliminate bias If the experiment corresponding to the data vector has a special experimental environment or low reliability, delete the data in the database; repeat (1.6) after screening the data that cause the funnel plot to be asymmetric, and recheck the publication bias; (1.7) Evaluate whether there is a general pattern in the data in the database; the data significance level is calculated as: Data significance tendency level F n When F is greater than zero, the data in the database are considered to have a significant general law. n <0, it is necessary to return to step (1.1) and add intervention measures to the article selection again.

4. The method for predicting agricultural mulch film degradation cycle based on RF-Meta model according to claim 1, It is characterized in that Step (2) is specifically as follows: (2.1) Scan the degradation rate data and variance data to check whether there are missing or null values. Once a null value is found, the location information of the null value needs to be recorded, including the row m and column index n where the null value is located, and fill it using the nearest linear interpolation method; (2.2) A multivariate random effects model for Meta-analysis was established for the database; the basic random effects part of the model was generated by the different literature sources from which the data were extracted; the formula of the multivariate random effects model was: yi=Xiβ+ui+∈i Where: yi is the effect size of the ith study, Xi is the covariate / predictor variable associated with the ith study, β is the regression coefficient of the model, ui is the random effect of the ith study, representing the heterogeneity between studies, and ∈i is the error term of the model; (2.3) The weight of each feature on the result is calculated by the following formula: Qm is the weight used to measure the impact of each feature on the result. j represents the Qm calculated for the jth feature, and k represents the number of data in the database that contains this feature. is the weight of the jth feature factor, x ·j is the degradation rate of the jth characteristic factor.

5. The method for predicting agricultural mulch film degradation cycle based on RF-Meta model according to claim 2, It is characterized in that The step (3) is specifically: 1) Initial state: Some data values ​​in the original database established by the Meta-analysis method are empty, and the jth characteristic data of the i-th data in the database is recorded as A ij ; 2) Random forest architecture: The data is organized according to the different types of feature data to be collected in step (1.3). Each feature category has different feature data. For each feature category, an independent decision tree model is established for each feature in it. At the same time, s features from different feature categories are randomly extracted and d independent decision tree models are established to simulate the mutual influence of features under different feature categories. Finally, a total of d+6 decision tree models are established to form a random forest model. 3) Extracting data relationship features: Each decision tree will judge and classify each feature j, and learn the data relationship under a specific feature classification from the data; the data relationship features include the value range of the feature, the association between the features, and the data pattern; 4) Progressive filling: traverse all the missing values ​​in the database and record the number of missing values ​​under different feature j columns as θ j , establish the matrix [θ 1 ,θ 2 ,…θ j ], and check whether there is a value less than or equal to 0. If there is, remove the data; then use the bubble sort method to sort the θ values ​​in the matrix from small to large; then fill all the data under the j feature in order from small to large according to the sorted results; finally traverse the matrix [θ 1 ,θ 2 ,…θ j ], traverse the data in the database again to ensure that there are no missing values; 5) Regression prediction: Based on the experimental group data with complete feature values ​​in the original database, the random forest algorithm is used. The d+6 decision trees established in 2) are used to calculate the values ​​judged by the d+6 decision trees respectively, and the average is taken to fill the missing values ​​in sequence.

6. The method for predicting agricultural mulch film degradation cycle based on RF-Meta model according to claim 1, It is characterized in that In step (4), the complete data set is divided into a modeling set and a validation set by random sampling.

7. The method for predicting agricultural mulch film degradation cycle based on RF-Meta model according to claim 1, It is characterized in that RF-Meta model optimized by five-fold cross validation.

8. The method for predicting agricultural mulch film degradation cycle based on RF-Meta model according to claim 1, It is characterized in that The hyperparameter optimization uses GridSearchCV to perform grid multi-core acceleration to search for the optimal hyperparameters.

9. An agricultural mulch degradation cycle prediction device based on RF-Meta model, comprising a processor and an executable program stored in a memory and executable on the processor, Features: When the processor executes the executable program, the method according to any one of claims 1 to 8 is implemented.

10. A storage medium containing a computer executable program, It is characterized in that When the computer executable program is executed by a computer processor, it is used to perform the method according to any one of claims 1 to 8.

Citation Information

Patent Citations

  • Prediction method, device and equipment for residual quantity of mulching film and storage medium

    CN114821371A

  • Method for predicting service life of biodegradable mulching film

    CN115711845A

  • Prediction method for pesticide adsorption distribution coefficient in soil with addition of agricultural organic materials

    CN116759011A

  • Birth defect risk factor assessment method based on causal learning

    CN116978556A

Cited By

  • Building industry carbon emission prediction method and system based on random forest algorithm

    CN119250845A

  • Construction industry carbon emission prediction method and system based on random forest algorithm

    CN119250845B

  • Modeling and analysis method and system for correlation between glaze wear resistance and process parameters

    CN119783557A

  • Multi-periodic ecological environment index anomaly identification and deletion interpolation method

    CN120724360A

  • Method for prolonging service life of biodegradable mulching film material

    CN120850798A