A method and system for evaluating the suitability of a low-yield field soil obstacle reduction technology

Abstract

The present application relates to the field of wisdom agriculture, and provides a kind of low yield field soil obstacle reduction technology suitability evaluation method and system.The method comprises: integrating field sampling data and public data to obtain a preliminary index pool;The preliminary index pool is input into the principal component analysis model, and the index dimensionality reduction is carried out to obtain a core evaluation index set;Based on the random effect model, the effect data of multiple reduction technologies in the public data on multiple indexes in the core evaluation index set are analyzed to obtain the comprehensive effect value and the corresponding confidence interval of the multiple technology index pairs;The decision matrix is constructed with the comprehensive effect value as the matrix element, the combination weighting processing and Monte Carlo uncertainty simulation are introduced into the TOPSIS decision model, and the relative closeness of multiple reduction technologies is calculated and obtained;According to the relative closeness, multiple reduction technologies are sorted in descending order to obtain the suitability recommendation result of soil obstacle reduction technology.The present application improves the objectivity of reduction technology suitability evaluation.

Description

Technical Field

[0001] This invention relates to the field of smart agriculture technology, and in particular to a method and system for evaluating the suitability of soil obstacle reduction technology in low- and medium-yield fields. Background Technology

[0002] Low- and medium-yield farmland commonly suffers from various soil obstacles such as salinization, acidification, infertility, compaction, and shallow topsoil, which severely restrict crop root development, water and fertilizer use efficiency, and the normal functioning of the farmland ecosystem. Scientifically evaluating the comprehensive effects of various obstacle reduction technologies is a crucial step in achieving targeted improvement and efficient resource allocation in low- and medium-yield farmland.

[0003] Existing evaluation methods for soil obstacle reduction technologies in low- and medium-yield fields generally suffer from the following shortcomings: the number of initial evaluation indicators is enormous and the indicators are highly correlated, making it easy for multicollinearity to distort the weight allocation when directly included in the evaluation model; at the same time, existing evaluation methods mostly rely on qualitative descriptions of single experiments or local areas, lacking a systematic integration of existing research results at the global or regional level, failing to form a statistically significant comprehensive effect size and objectively quantify its uncertainty, resulting in fragmented and incomparable technical effect assessment results, making it difficult to provide a reliable decision-making basis for the screening and promotion of cross-regional soil obstacle reduction technologies. Summary of the Invention

[0004] This invention provides a method and system for evaluating the suitability of soil obstacle reduction technology in low- and medium-yield fields, in order to overcome the shortcomings of existing technologies.

[0005] The first aspect of this invention provides a method for evaluating the suitability of soil obstacle reduction technologies in low- and medium-yield farmland, comprising: S1: Integrate the field sampling data and publicly available data of the target area to obtain a preliminary index pool containing multiple soil barrier-related indicators; S2: Input the initial index pool into the principal component analysis model to perform index dimensionality reduction and obtain the core evaluation index set; S3: Based on the random effects model, the effect data of multiple reduction techniques on multiple indicators in the core evaluation indicator set in the public data are analyzed to obtain the comprehensive effect value and corresponding confidence interval of multiple technical indicator pairs; S4: Construct a decision matrix using the combined effect value as matrix elements, introduce combined weighting processing and Monte Carlo uncertainty simulation into the TOPSIS decision model, and calculate the relative closeness of multiple reduction techniques. S5: Soil obstacle reduction technologies are sorted in descending order based on their relative proximity to obtain suitability recommendations for soil obstacle reduction technologies.

[0006] According to the suitability evaluation method for soil obstacle reduction technology in low- and medium-yield fields provided by the present invention, step S1 further includes: S11: Systematically set up sampling points in the target area, collect topsoil samples, and systematically measure the physical, chemical and biological properties of the soil to obtain field sampling data. S12: Based on the soil obstacle type in the target area, a systematic search of public databases is conducted according to the preset inclusion criteria to obtain a literature dataset; S13: Integrate the field sampling data with the literature dataset, and conduct preliminary screening of candidate indicators based on soil barrier mechanisms to obtain a preliminary indicator pool.

[0007] According to the suitability evaluation method for soil obstacle reduction technology in low- and medium-yield farmland provided by the present invention, in step S12, the preset inclusion criteria include: Both experimental and control groups were set up. Core statistical information includes available or convertible sample size, mean, and standard deviation; The number of experiments repeated shall not be less than the preset number and the number of soil samples shall not be less than the preset quantity; The depth of the soil layer was uniformly limited to a preset depth range. The measurement units and standards for multiple indicators are consistent.

[0008] According to the suitability evaluation method for soil obstacle reduction technology in low- and medium-yield fields provided by the present invention, step S2 further includes: S21: Standardize the sample data of the initial index pool and calculate the covariance matrix of the standardized data; S22: Perform eigenvalue decomposition on the covariance matrix, extract multiple principal components according to the cumulative variance contribution rate, and obtain the principal component loading matrix; S23: Analyze the principal component loading matrix, select the representative index with the largest absolute value of the loading from multiple principal components, merge the representative indices corresponding to multiple principal components to remove duplicates, and obtain the core evaluation index set.

[0009] According to the suitability evaluation method for soil obstacle reduction technology in low- and medium-yield fields provided by the present invention, step S3 further includes: S31: Extract statistical data of the experimental and control groups of multiple indicators in the core evaluation index set from the publicly available dataset using multiple reduction techniques; S32: Based on the statistical data of the experimental and control groups for multiple indicators, the logarithm of the ratio of the experimental group mean to the control group mean is used as the logarithmic response ratio. The logarithmic response ratio and corresponding variance of multiple study records are calculated to obtain the effect size dataset of multiple study records. S33: Input the effect size dataset into the random effects model, merge multiple records of the same technical indicator pair, and convert the merged log response ratio into a percentage effect value to obtain the comprehensive effect value; S34: Determine the corresponding confidence interval based on the standard error corresponding to the comprehensive effect value, and determine the statistical significance of multiple technical indicators on the effect through significance testing.

[0010] According to the suitability evaluation method for soil obstacle reduction technology in low- and medium-yield fields provided by the present invention, step S4 further includes: S41: Using the comprehensive effect value as matrix elements, with the reduction techniques to be evaluated as rows and multiple indicators from the core evaluation indicator set as columns, construct a basic decision matrix; S42: Standardize the basic decision matrix according to the attribute types of multiple indicators to obtain a standardized matrix; S43: The standardized matrix is ​​weighted using the entropy weighting method and the analytic hierarchy process, and the resulting objective weights and subjective weights are linearly fused to obtain the combined weights. S44: Multiply the standardized matrix by the combined weights to obtain the weighted decision matrix; S45: Using the confidence interval corresponding to the comprehensive effect value as the sampling range, perform Monte Carlo random sampling on the basic decision matrix to generate a preset number of simulated decision matrices, and perform combined weighting processing on multiple simulated decision matrices to obtain a corresponding number of weighted simulated decision matrices. S46: Determine the positive and negative ideal solutions for each index on multiple weighted simulation decision matrices, calculate the Euclidean distances between multiple reduction techniques and the positive and negative ideal solutions, and obtain the relative closeness of multiple reduction techniques.

[0011] According to the suitability evaluation method for soil obstacle reduction technology in low- and medium-yield fields provided by the present invention, in step S5, the recommendation level of the suitability recommendation result includes: priority recommendation, recommendation, optional, and not recommended.

[0012] A second aspect of the present invention provides a suitability evaluation system for soil obstacle reduction technology in low- and medium-yield fields, comprising: The data acquisition module is used to integrate field sampling data and publicly available data from the target area to obtain a preliminary index pool containing multiple soil obstacle-related indicators. Dimensionality reduction module: used to input the initial index pool into the principal component analysis model, perform index dimensionality reduction, and obtain the core evaluation index set; Analysis module: used to analyze the effect data of multiple reduction techniques on multiple indicators in the core evaluation indicator set based on the random effects model, and to obtain the comprehensive effect value and corresponding confidence interval of multiple technical indicator pairs; The calculation module is used to construct a decision matrix with the combined effect value as matrix elements, introduce the combined weighting process and Monte Carlo uncertainty simulation into the TOPSIS decision model, and calculate the relative closeness of multiple reduction techniques. Output module: Used to sort multiple mitigation techniques in descending order based on their relative proximity, and obtain the suitability recommendation results for soil obstacle reduction techniques.

[0013] A third aspect of the present invention also provides a suitability evaluation device for soil obstacle reduction technology in low- and medium-yield fields, comprising: a memory and at least one processor, wherein the memory stores instructions; at least one processor invokes the instructions in the memory to cause the suitability evaluation device for soil obstacle reduction technology in low- and medium-yield fields to perform a suitability evaluation method for soil obstacle reduction technology in low- and medium-yield fields as described in any of the preceding claims.

[0014] A fourth aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed by a processor, implement a suitability evaluation method for soil obstacle reduction technology in low- and medium-yield farmland as described in any of the preceding claims.

[0015] This invention provides a method, system, equipment, and storage medium for evaluating the suitability of soil obstacle reduction technologies in low- and medium-yield farmland. By organically coupling principal component analysis, meta-analysis, and the TOPSIS decision model, and introducing Monte Carlo uncertainty simulation and combined weighting mechanisms, it achieves breakthroughs over existing evaluation methods at multiple levels. At the indicator selection level, the introduction of principal component analysis allows the evaluation indicator system to extract the core evaluation indicator set that truly dominates soil obstacles from a redundant initial indicator pool, fundamentally eliminating collinearity interference between indicators. This makes subsequent model inputs more refined and weight allocation more reasonable, effectively reducing evaluation costs and time. At the effect quantification level, meta-analysis based on a random-effects model systematically integrates global or regional literature data into a statistically significant comprehensive effect value, and simultaneously outputs confidence intervals to objectively characterize the uncertainty of effect estimation. This ensures that the technical effect knowledge on which the evaluation is based is established on systematic scientific evidence, rather than being limited to single experimental conclusions, greatly improving the universality and credibility of the evaluation conclusions. At the decision-making level, this invention, through the linear fusion of entropy weighting and the analytic hierarchy process (AHP), preserves the objective differences revealed by the data itself while incorporating prior domain knowledge under specific obstacle scenarios, making the weight allocation both objective and situationally adaptable. The introduction of Monte Carlo simulation transforms the confidence interval of the comprehensive effect value into the random fluctuation of the decision matrix, ensuring that the final relative closeness calculation fully considers the uncertainty of effect estimation, resulting in more robust and reliable evaluation conclusions. Furthermore, this invention uses public databases as its knowledge source, allowing for continuous updates to the meta-analysis results as public data is continuously incorporated, enabling the evaluation system to continuously self-optimize. This allows for rapid ranking and selection of reduction techniques without the need for large-scale field trials, significantly saving decision-making costs. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of a suitability evaluation method for soil obstacle reduction technology in low- and medium-yield fields provided by the present invention. Figure 2 PCA gravel map provided for embodiments of the present invention; Figure 3 Principal component loading diagrams provided for embodiments of the present invention; Figure 4 Forest plot for meta-analysis of technical indicator pairs provided in embodiments of the present invention; Figure 5 This is a schematic diagram of the TOPSIS model calculation principle provided in an embodiment of the present invention; Figure 6 A ranking chart of recommendation results provided for embodiments of the present invention; Figure 7 This is a schematic diagram of the structure of a suitability evaluation system for soil obstacle reduction technology in low- and medium-yield fields provided by the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, embodiments of this invention, and should not be construed as limiting the invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention. In the description of this invention, it should be understood that the terminology used is for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0019] The embodiments of the present invention are described below with reference to the figures.

[0020] like Figure 1 As shown, this invention provides a method for evaluating the suitability of soil obstacle reduction technology in low- and medium-yield fields, comprising: S1: Integrate the field sampling data and publicly available data of the target area to obtain a preliminary index pool containing multiple soil obstacle-related indicators.

[0021] Step S1 further includes: S11: Systematically set up sampling points in the target area, collect topsoil samples, and systematically measure the physical, chemical and biological properties of the soil to obtain field sampling data.

[0022] In step S11, the present invention first systematically sets up sampling points within the target area to collect soil samples from the 0–20 cm topsoil layer. Subsequently, for each sampling point, the present invention measures three categories of soil property indicators: physical property indicators, including parameters reflecting soil structure such as water-stable aggregates, topsoil thickness, and bulk density; chemical property indicators, including parameters reflecting soil nutrient supply and pH status such as pH, organic matter, total nitrogen, total phosphorus, available nitrogen, available phosphorus, available potassium, and cation exchange capacity; and biological property indicators, including parameters reflecting soil biological activity such as enzyme activity, microbial biomass carbon, microbial biomass nitrogen, and microbial diversity. After collection, the measurement results of the three categories of indicators for each sampling point are organized according to the sampling point number to form a field sampling data table with the sampling points as rows and the measured indicators as columns, i.e., field sampling data.

[0023] S12: Based on the soil obstacle type in the target area, a systematic search of public databases is conducted according to the preset inclusion criteria to obtain a literature dataset.

[0024] In step S12, the preset inclusion criteria include: setting up both an experimental group and a control group; the core statistical information includes the obtainable or convertible sample size, mean, and standard deviation; the number of experimental repetitions is not less than the preset number and the soil sample size is not less than the preset quantity; the depth of the research soil layer is uniformly limited to the preset depth range; and the measurement units and calibers of multiple indicators are uniform.

[0025] In step S12, after the field sampling data is acquired, the present invention systematically searches domestic and foreign databases according to preset inclusion criteria, focusing on the soil obstacle type of the target area and combining keywords such as soil quality improvement technology and improvement materials.

[0026] The preset inclusion criteria are a set of screening rules for included literature in terms of experimental design and data completeness, specifically including the following five items: First, the literature must simultaneously set up an experimental group and a control group. The experimental group is the treatment group that received the amendment or technical treatment, and the control group is the untreated or baseline treatment group. Both must be present for the effect value to be calculated. Second, the core statistical information must include the obtainable or convertible sample size, mean, and standard deviation. If the literature only reports the standard error, it should be converted according to the mathematical conversion relationship between standard deviation and standard error. Third, the number of experimental replicates and the soil sample size must both reach the preset lower limit to ensure the reliability of the statistical conclusions. Fourth, the soil depth studied must be uniformly limited to the preset depth range, with the 0–20cm topsoil layer as the uniform caliber. Research results from different soil depths are not comparable and are excluded. Fifth, the measurement units and measurement calibers of each indicator must be horizontally comparable. Literature that cannot be uniformly converted is eliminated. After screening according to the above five criteria, literature lacking a control group, with incomplete data, or with inconsistent indicator calibers is excluded, resulting in a literature dataset that meets the requirements.

[0027] S13: Integrate the field sampling data with the literature dataset, and conduct preliminary screening of candidate indicators based on soil barrier mechanisms to obtain a preliminary indicator pool.

[0028] In step S13, after acquiring the literature dataset and the field sampling data, this invention integrates the two. Specifically, the integration is based on the soil index categories measured in the field sampling data, comparing them with the indices involved in various studies in the literature dataset. Candidate indices that are recorded in both the field sampling data and the literature dataset, and whose indices have consistent definitions, are selected to form a candidate index set. Subsequently, this invention performs preliminary screening of the above candidate indices based on soil obstacle mechanisms. "Based on soil obstacle mechanisms" means determining whether each candidate index is directly related to the formation, evolution, and improvement effect of the target obstacle type (such as infertility, acidification, salinization, etc.) according to its causal mechanism. Only indices with mechanistic correlation to the obstacle type are retained, while redundant parameters without substantial correlation to the target obstacle are excluded, ultimately forming a preliminary index pool.

[0029] S2: Input the initial index pool into the principal component analysis model to perform index dimensionality reduction and obtain the core evaluation index set.

[0030] Step S2 further includes: S21: Standardize the sample data of the initial index pool and calculate the covariance matrix of the standardized data.

[0031] Furthermore, the initial index pool contains multiple soil barrier-related indicators, and these indicators have significant differences in dimensions and orders of magnitude. If the covariance matrix is ​​directly calculated, the indicators with larger dimensions will numerically dominate the subsequent calculation results, leading to biases in principal component extraction. Therefore, this invention first performs Z-score standardization on all sample data in the initial index pool.

[0032] Specifically, this invention uses Z-score standardization, subtracting the mean of each indicator from all samples and then dividing by the standard deviation of that indicator. After processing, the mean of each indicator is 0 and the standard deviation is 1, thus unifying the numerical scale of each indicator and obtaining a standardized data matrix. Subsequently, using the standardized data matrix as input, the covariance between each pair of indicators is calculated to form a covariance matrix.

[0033] The obtained covariance matrix contains the value of each element, which reflects the degree of linear correlation between the two corresponding indicators. Positive values ​​indicate that they change in the same direction, while negative values ​​indicate that they change in opposite directions. The larger the absolute value, the stronger the linear relationship between the two indicators. This matrix is ​​the mathematical basis for subsequent eigenvalue decomposition and principal component extraction.

[0034] S22: Perform eigenvalue decomposition on the covariance matrix, extract multiple principal components according to the cumulative variance contribution rate, and obtain the principal component loading matrix.

[0035] In step S22, after obtaining the covariance matrix, this invention performs eigenvalue decomposition. Eigenvalue decomposition is the process in linear algebra of decomposing a matrix into eigenvalues ​​and eigenvectors. The magnitude of the eigenvalue represents the proportion of the total variance of the original data that the corresponding principal component can explain, while the eigenvector determines which original indicators are linearly combined with which coefficients to form the principal component. After decomposition, the principal components are arranged in descending order of their corresponding eigenvalues, and the variance contribution rate (i.e., the proportion of the principal component's eigenvalue to the sum of all eigenvalues) and cumulative variance contribution rate of each principal component are calculated sequentially. This invention uses a preset threshold for the cumulative variance contribution rate as a criterion to extract the first few principal components, ensuring that the extracted principal components can fully represent the overall variance information of the original indicator pool. Finally, the eigenvectors corresponding to each principal component are arranged column-wise to form a principal component loading matrix. Each column in the loading matrix corresponds to a principal component, and each row corresponds to an original indicator. The matrix elements are the loading values ​​of the original indicator on the corresponding principal component. The larger the absolute value of the loading value, the greater the contribution of the original indicator to the principal component.

[0036] S23: Analyze the principal component loading matrix, select the representative index with the largest absolute value of the loading from multiple principal components, merge the representative indices corresponding to multiple principal components to remove duplicates, and obtain the core evaluation index set.

[0037] In step S23, after obtaining the principal component loading matrix, this invention analyzes the loading columns corresponding to each principal component one by one. From each principal component, the original index with the largest absolute value of the loading is selected as the representative index of that principal component. This representative index has the strongest explanatory power in the direction of variation captured by that principal component. Subsequently, this invention summarizes the representative indices selected by each principal component and performs a merge and deduplication process. That is, if the same original index is selected as a representative index in multiple principal components, it is only retained once during the summary. The index set obtained after merging and deduplication is the core evaluation index set. Each index in the core evaluation index set corresponds to the direction of variation represented by different principal components. The collinearity between the indices is significantly lower than that in the initial index pool, and each index has the strongest representativeness on a certain principal component, thus covering the main variation information of the initial index pool with a smaller number of indices.

[0038] S3: Based on the random effects model, the effect data of multiple reduction techniques on multiple indicators in the core evaluation indicator set are analyzed to obtain the comprehensive effect value and corresponding confidence interval of multiple technical indicator pairs.

[0039] Step S3 further includes: S31: Extract statistical data of the experimental and control groups of multiple indicators in the core evaluation index set from the publicly available dataset using multiple reduction techniques.

[0040] In step S31, after the core evaluation index set is determined, this invention extracts data selectively from the literature dataset. Specifically, this invention uses each index in the core evaluation index set as the retrieval object, traversing each article in the literature dataset one by one, and extracting statistical data of the experimental and control groups for each reduction technology (such as the application of biochar, organic fertilizer, soil conditioner, etc.) on each index. Each extracted record contains the following information: reduction technology name, corresponding index name, experimental group sample size, experimental group mean, experimental group standard deviation, control group sample size, control group mean, and control group standard deviation. The above information constitutes the original input for subsequent effect size calculation. If a published article only reports the standard error instead of the standard deviation, this invention converts it according to the conversion relationship between standard deviation and standard error before including it. After extracting and organizing each article, a structured data table is finally formed with each processed data as a row and the above statistical information as columns, serving as the data basis for effect size calculation.

[0041] S32: Based on the statistical data of the experimental and control groups for multiple indicators, the logarithm of the ratio of the experimental group mean to the control group mean is used as the logarithmic response ratio. The logarithmic response ratio and corresponding variance of multiple study records are calculated to obtain the effect size dataset of multiple study records.

[0042] After extracting the statistical data of the experimental and control groups for each study record, step S32 of this invention aims to calculate the effect size for each study record. This invention uses the logarithmic response ratio as the effect size indicator. The logarithmic response ratio is a commonly used measure of effect size for continuous variables in meta-analysis, expressed as the natural logarithm of the ratio of the experimental group mean to the control group mean. Its positive and negative values ​​reflect the increasing or decreasing trend of the experimental group relative to the control group, respectively, while the absolute value reflects the effect strength. Furthermore, the logarithmic transformation makes the distribution of the effect size closer to a normal distribution, which is beneficial for subsequent statistical aggregation.

[0043] Specifically, this invention calculates the log-response ratio for each research record by taking the natural logarithm of the ratio of the experimental group mean to the control group mean. Then, based on the means, standard deviations, and sample sizes of the experimental and control groups, the variance of the log-response ratio for that record is calculated. The magnitude of the variance reflects the accuracy of the effect size estimate; a smaller variance indicates a more accurate estimate of the true effect, resulting in a greater weight in subsequent weighted aggregation. Finally, the log-response ratios and corresponding variances of all research records are compiled and summarized to obtain the effect size dataset.

[0044] Furthermore, the logarithmic response ratio in step S32 The calculation formula is:

[0045] Log response ratio variance The calculation formula is:

[0046] in, This is the mean of the experimental group. This is the mean of the control group. The standard deviation of the experimental group is 1. The standard deviation of the control group is 1. The sample size of the experimental group. This represents the sample size of the control group.

[0047] S33: Input the effect size dataset into the random effects model, merge multiple records of the same technical indicator pair, and convert the merged log response ratio into a percentage effect value to obtain the combined effect value.

[0048] In step S33, after the effect size dataset is constructed, this invention inputs it into a random effects model to perform a weighted merging operation on multiple study records under the same technology-indicator pair. Furthermore, the random effects model described in this invention has the advantage of allowing for heterogeneity of the true effect among different studies. By estimating the variance between studies in addition to the within-study variance, the weight of each record is adjusted, making the estimation of the combined effect value more robust under different experimental conditions, soil background, and climate backgrounds. Specifically, the merging weight of each study record is determined by the reciprocal of its variance and the variance between studies; records with smaller variances (i.e., higher estimation accuracy) receive larger weights. After weighted merging, the merged logarithmic response ratio of the technology-indicator pair is obtained. Subsequently, this invention performs an exponential transformation on the merged logarithmic response ratio, subtracts 1, and multiplies by 100 to convert it into a percentage effect value, i.e., the combined effect value, which intuitively reflects the percentage change in the corresponding indicator compared to the control group caused by the reduction technique.

[0049] S34: Determine the corresponding confidence interval based on the standard error corresponding to the comprehensive effect value, and determine the statistical significance of multiple technical indicators on the effect through significance testing.

[0050] After determining the combined effect value, step S34 of this invention further quantifies its statistical precision and significance. Regarding precision quantification, this invention approximates the standard error of the pooled log-response ratio as the standard error of the percentage effect value using a first-order Taylor expansion. This standard error is then symmetrically expanded to both sides to determine the confidence interval corresponding to the combined effect value. The width of the confidence interval directly reflects the consistency of effect estimation across studies; a narrower interval indicates a more concentrated effect estimation among different studies and a more reliable overall conclusion. If the confidence interval does not contain zero, it initially indicates that the effect direction of this technique on the indicator is clear. Regarding significance determination, this invention obtains the Z-statistic by dividing the pooled log-response ratio by its standard error. The Z-statistic is then substituted into the cumulative distribution function of the standard normal distribution to calculate the two-sided P-value. Whether the P-value is lower than a preset significance level determines the statistical significance of the effect of the technique-indicator pair. After the above processing, each technique-indicator pair ultimately corresponds to a combined effect value, a confidence interval, and a significance determination result. These three together constitute the effect size information required for subsequent construction of the decision matrix.

[0051] In step S34, the formula for calculating the standard error of the combined effect value is:

[0052] in, The combined effect value The standard error, The standard error of the log-response ratio after pooling random effects models. To combine the exponential transformation values ​​of the logarithmic response ratio, the confidence interval is... Add or subtract respectively Obtain the upper and lower bounds.

[0053] S4: Construct a decision matrix using the combined effect value as matrix elements, and introduce combined weighting processing and Monte Carlo uncertainty simulation into the TOPSIS decision model to calculate the relative proximity of multiple reduction techniques.

[0054] Step S4 further includes: S41: Using the comprehensive effect value as matrix elements, with the reduction techniques to be evaluated as rows and multiple indicators from the core evaluation indicator set as columns, construct a basic decision matrix.

[0055] For each technology-indicator pair obtained in step S3, corresponding to a comprehensive effect value, this invention constructs a basic decision matrix using these comprehensive effect values ​​as matrix elements. The rows of the basic decision matrix correspond to each reduction technology to be evaluated, and the columns correspond to each indicator in the core evaluation indicator set. The element in the i-th row and j-th column of the matrix is ​​the comprehensive effect value of the i-th reduction technology on the j-th indicator. The comprehensive effect value is expressed as a percentage, representing the magnitude of change in the corresponding indicator compared to the control group. A positive value indicates that the technology improves the indicator relative to the control group, and a negative value indicates a decrease. Through the above arrangement, the effect size information of each reduction technology on each indicator is uniformly organized into a structured matrix, namely the basic decision matrix, which serves as the original data carrier for all subsequent decision calculations.

[0056] S42: Standardize the basic decision matrix according to the attribute types of multiple indicators to obtain a standardized matrix.

[0057] After the basic decision matrix is ​​constructed, since the comprehensive effect values ​​of each indicator differ in numerical range and there is no direct numerical comparability between different indicators, the basic decision matrix needs to be standardized. Therefore, this invention adopts different standardization methods according to the attribute type of each indicator.

[0058] For positive indicators, i.e., those with larger effect values ​​(such as soil nutrient enhancement indicators like organic matter, available potassium, available nitrogen, and available phosphorus), this invention employs an extreme value standardization method. The minimum value of the indicator is subtracted from the comprehensive effect value of each reduction technology on that indicator, and then divided by the difference between the maximum and minimum values. This linearly compresses the original effect value to a range of 0 to 1, with larger values ​​indicating better performance of the technology on that indicator. For moderate indicators, i.e., those with an optimal range for effect values, where deviation from this range leads to poorer performance, this invention uses a preset optimal range as a benchmark for range transformation. The degree of deviation of each comprehensive effect value from the optimal range is converted into a standardized score of 0 to 1. The effect value falling within the optimal range receives the highest score, and the greater the deviation, the lower the score.

[0059] After the above classification process, the standardized values ​​of each indicator fall within the range of 0 to 1. The standardized values ​​of each technology on different indicators are horizontally comparable. The standardized values ​​of each indicator are rearranged according to the original matrix structure to obtain the standardized matrix.

[0060] S43: The standardized matrix is ​​weighted using the entropy weighting method and the analytic hierarchy process, and the resulting objective weights and subjective weights are linearly fused to obtain the combined weights.

[0061] After the standardized matrix is ​​determined, this invention assigns weights to each indicator to reflect the relative importance of different indicators in the comprehensive evaluation. This invention employs both the entropy weighting method and the analytic hierarchy process (AHP) to obtain objective and subjective weights respectively, and then linearly merges the two into a combined weight.

[0062] Entropy weighting is an objective weighting method that determines weights entirely based on data distribution characteristics. This invention uses a standardized matrix as input. First, it calculates the proportion of the standardized value of each reduction technique under each indicator to the sum of the standardized values ​​of all techniques for that indicator, obtaining the proportion matrix of each technique for each indicator. Then, based on the proportion matrix, it calculates the information entropy of each indicator. Information entropy reflects the uniformity of the indicator's distribution among all evaluated techniques. The closer the standardized values ​​of each technique are to the indicator, the larger the entropy value, indicating that the indicator contributes less to distinguishing different techniques. Based on the information entropy, it calculates the difference coefficient for each indicator. The difference coefficient equals 1 minus the information entropy of the corresponding indicator. The larger the difference coefficient, the more significant the distribution difference of the indicator among different techniques. Finally, it normalizes the difference coefficients of each indicator so that the sum of the difference coefficients of all indicators is 1, obtaining the objective entropy weight of each indicator.

[0063] The Analytic Hierarchy Process (AHP) is a subjective weighting method that relies on domain expert knowledge to determine weights. This invention addresses soil obstacle scenarios in a target area. Domain experts assess the relative importance of each indicator in a core evaluation index set, assigning a subjective importance score to each indicator, thus constructing a subjective importance vector. Subsequently, a consensus judgment matrix is ​​constructed using the ratio of the subjective importance scores of any two indicators as elements. The element in the j-th row and k-th column of this matrix represents the ratio of the importance of the j-th indicator to the k-th indicator. The eigenvector corresponding to the largest eigenvalue of this consensus judgment matrix is ​​solved, and this eigenvector is normalized to obtain the subjective weights of the AHP.

[0064] After obtaining the objective weight of entropy weight and the subjective weight of analytic hierarchy process, this invention linearly fuses the two according to a preset fusion coefficient. The fusion coefficient determines the proportion of each objective and subjective weight in the combined weight, thus obtaining the combined weight of each indicator.

[0065] S44: Multiply the standardized matrix by the combined weights to obtain the weighted decision matrix.

[0066] In step S44, after determining the combined weights, this invention multiplies the standardized matrix by the combined weights to obtain a weighted decision matrix. Specifically, each column of the standardized matrix (i.e., the standardized value of all technologies corresponding to each indicator) is multiplied by the combined weight corresponding to that indicator. This scales the standardized value of each indicator according to its combined weight, amplifying the numerical contribution of more important indicators in the matrix and reducing the contribution of less important indicators. After column-wise weighting, the sum of the values ​​in each column is no longer equal, and the numerical influence of each indicator in the matrix directly corresponds to its combined weight, resulting in the weighted decision matrix. The element in the i-th row and j-th column of the weighted decision matrix is ​​the weighted standardized effect value of the i-th reduction technology on the j-th indicator, serving as the direct input for subsequent TOPSIS distance calculation.

[0067] S45: Using the confidence interval corresponding to the comprehensive effect value as the sampling range, perform Monte Carlo random sampling on the basic decision matrix to generate a preset number of simulated decision matrices, and perform combined weighting processing on multiple simulated decision matrices to obtain a corresponding number of weighted simulated decision matrices.

[0068] Steps S41 to S44 construct a weighted decision matrix based on point estimates of the combined effect value. In step S3, each combined effect value corresponds to a confidence interval, reflecting a certain degree of uncertainty in the effect size estimate. To incorporate this uncertainty into the decision-making process, Monte Carlo random sampling is introduced in step S45 of this invention. This aims to estimate the output distribution of a complex system through extensive random sampling simulations, and this invention applies it to the uncertainty propagation of the decision matrix.

[0069] Specifically, this invention performs a random sampling of each element in the basic decision matrix, centered on the corresponding comprehensive effect value and within the corresponding confidence interval, assuming the effect value follows a normal distribution within this interval, to obtain a random simulated value for that element. After simultaneously performing a random sampling of all elements in the basic decision matrix, a simulated decision matrix is ​​obtained. This random sampling process is repeated a predetermined number of times to obtain a corresponding number of simulated decision matrices, each representing a reasonable realization within the uncertainty range of the comprehensive effect value. Subsequently, each simulated decision matrix is ​​processed one by one according to the standardization and weighting process in steps S42 to S44 to obtain a weighted simulated decision matrix corresponding to the number of simulated decision matrices.

[0070] S46: Determine the positive and negative ideal solutions for each index on multiple weighted simulation decision matrices, calculate the Euclidean distances between multiple reduction techniques and the positive and negative ideal solutions, and obtain the relative closeness of multiple reduction techniques.

[0071] After obtaining all weighted simulation decision matrices, the present invention performs TOPSIS calculation on each weighted simulation decision matrix. That is, in the multi-index space, the positive ideal solution and negative ideal solution of each index are determined, and the distance between each object to be evaluated and the positive and negative ideal solutions is calculated. The closer the object is to the positive ideal solution and the farther the object is from the negative ideal solution, the better the object is. The relative proximity of each object to be evaluated is calculated to reflect the relative superiority or inferiority of its overall performance.

[0072] Specifically, for each weighted simulation decision matrix, this invention scans each column sequentially, determining the maximum value of the weighted standardized value corresponding to all reduction techniques in each index column as the positive ideal solution for that index, and the minimum value as the negative ideal solution. The positive and negative ideal solution vectors represent the best and worst possible levels achievable for each index in the current simulation, respectively. Subsequently, for each reduction technique, the arithmetic square root of the sum of the squares of the differences between the corresponding row vector and the positive ideal solution vector in the weighted simulation decision matrix is ​​calculated to obtain the Euclidean distance between the technique and the positive ideal solution; similarly, the Euclidean distance between the technique and the negative ideal solution is calculated.

[0073] After both Euclidean distances are determined, the present invention divides the Euclidean distance between the technology and the negative ideal solution by the sum of the Euclidean distances between the technology and the positive ideal solution and the negative ideal solution to obtain the relative closeness of the technology in the current simulation. The value range of the relative closeness is 0 to 1. The larger the value, the closer the overall performance of the technology in the current simulation is to the positive ideal solution.

[0074] After performing the above TOPSIS calculations for all preset simulations, each reduction technique obtains a relative closeness sample equal to the number of simulations. This invention takes the average of the relative closeness of each reduction technique across all simulations as the final relative closeness of the technique.

[0075] S5: Soil obstacle reduction technologies are sorted in descending order based on their relative proximity to obtain suitability recommendations for soil obstacle reduction technologies; wherein the recommendation levels of the suitability recommendations include: preferred, recommended, optional, and not recommended.

[0076] Furthermore, after determining the final relative closeness of all mitigation techniques, this invention ranks each technique in descending order of relative closeness, with higher-ranking techniques showing a closer approximation to the positive ideal solution in terms of comprehensive performance across multiple indicators. Subsequently, using a preset relative closeness threshold as a boundary, this invention divides the ranking results into four recommendation levels: the technique with the highest relative closeness is designated as the priority recommendation level, followed by the recommendation level, then the optional level, and the technique with the lowest relative closeness is designated as the non-recommended level. Taking the soil infertility obstacle scenario in North China as an example, a relative closeness of 0.55 is used as the boundary threshold between priority and recommendation, and 0.45 is used as the boundary threshold between recommendation and optional. Techniques below 0.45 are further categorized as optional or non-recommended based on their specific values. The recommendation level of each technique, along with its corresponding relative closeness and the Euclidean distance between the positive and negative ideal solutions, is output together to form a suitable recommendation list for soil obstacle mitigation techniques.

[0077] like Figure 2 As shown, a second aspect of the present invention provides a suitability evaluation system for soil obstacle reduction technology in low- and medium-yield farmland, comprising: Data Acquisition Module 100: Used to integrate field sampling data and publicly available data of the target area to obtain a preliminary index pool containing multiple soil obstacle-related indicators; Dimensionality reduction module 200: used to input the initial index pool into the principal component analysis model, perform index dimensionality reduction, and obtain the core evaluation index set; Analysis module 300: is used to analyze the effect data of multiple reduction techniques on multiple indicators in the core evaluation indicator set based on the random effects model, and to obtain the comprehensive effect value and corresponding confidence interval of multiple technical indicator pairs; Calculation module 400: Used to construct a decision matrix with the combined effect value as matrix element, introduce combined weighting processing and Monte Carlo uncertainty simulation into the TOPSIS decision model, and calculate the relative closeness of multiple reduction techniques; Output module 500: Used to sort multiple mitigation technologies in descending order based on their relative proximity, and obtain the suitability recommendation results for soil obstacle reduction technologies.

[0078] The present invention also provides a suitability evaluation device for soil obstacle reduction technology in low- and medium-yield fields, comprising: a memory and at least one processor, wherein the memory stores instructions; at least one processor invokes the instructions in the memory to cause the suitability evaluation device for soil obstacle reduction technology in low- and medium-yield fields to perform a suitability evaluation method for soil obstacle reduction technology in low- and medium-yield fields as described in any of the above.

[0079] The present invention also provides a computer-readable storage medium storing instructions that, when executed by a processor, implement a suitability evaluation method for soil obstacle reduction technology in low- and medium-yield farmland as described in any of the preceding claims.

[0080] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0081] The following specific embodiments further illustrate the suitability evaluation method for soil obstacle reduction technology in low- and medium-yield farmland according to the present invention. This embodiment takes the suitability evaluation of soil infertility obstacle reduction technology in a certain region as an example to explain the application process of the present invention in detail.

[0082] S101: Constructing a multi-source database and initial index pool. This embodiment selects soil infertility barriers in a certain region as the research object. During implementation, sampling points were systematically set up in the study area to collect soil samples from the 0–20 cm topsoil layer. A total of 17 indicators covering three major categories—physical, chemical, and biological—were systematically measured. The statistical results of each indicator are shown in Table 1.

[0083] Physical properties included water-stable aggregates, topsoil thickness, and bulk density; chemical properties included pH, organic matter, total nitrogen, total phosphorus, available nitrogen, available phosphorus, available potassium, nitrate nitrogen, ammonium nitrogen, and cation exchange capacity; biological properties included enzyme activity, microbial biomass carbon, microbial biomass nitrogen, and microbial diversity. Table 1 shows that all indicators exhibited some degree of spatial variability within the study area, with total phosphorus showing the largest coefficient of variation (34.8%), while available potassium showed a relatively smaller coefficient of variation (11.3%), reflecting the spatial heterogeneity of soil infertility barriers in North China in terms of nutrient content and soil structure.

[0084] Table 1. Measurement results of each index in this embodiment.

[0085] Regarding literature data acquisition, a systematic search and organization of relevant literature on soil infertility reduction technologies in the experimental area was conducted in Chinese and foreign literature databases based on preset search queries and inclusion screening criteria. Through evaluation in this embodiment, a total of 73 relevant articles were identified. 38 articles were excluded due to missing control groups, experimental groups, or incomplete data (missing sample size, standard deviation, or mean), resulting in 35 articles meeting the inclusion criteria. Of these, 21 were included in the heterogeneity analysis. Valid data were extracted from the eligible articles, and subsequent analysis was performed on 117 sets of processed data from these 25 articles, forming the literature dataset of this embodiment.

[0086] S102: Based on principal component analysis (PCA) for core indicator selection, PCA was performed on the sample data of 17 indicators in the initial indicator pool. First, the original data was Z-score standardized, and the covariance matrix of the standardized data was calculated. Then, eigenvalue decomposition was performed on the covariance matrix to obtain the variance contribution rate of each principal component. Combined with... Figure 2 As shown in the principal component scree plot, the variance contribution rate of the first principal component is 52.25%, the second principal component is 27.61%, the third principal component is 15.32%, and the fourth principal component is 4.00%. The cumulative variance contribution rate of the first three principal components reaches 95.19%, which exceeds the preset threshold. Therefore, the first three principal components are extracted to participate in the subsequent load analysis.

[0087] Combination such as Figure 3 The principal component loading plots shown analyze the loading matrices of each principal component. (From...) Figure 3 It can be seen that in the PC1 direction (contribution rate 52.26%), the absolute values ​​of the loadings of organic matter and available nitrogen are the largest. Both extend along the positive direction of PC1 and their directions are similar, indicating that organic matter and available nitrogen are the most representative in the direction of soil comprehensive nutrient variation represented by PC1. In the PC2 direction (contribution rate 27.61%), the absolute value of the loading of available potassium is the largest, and its vector extends along the negative direction of PC2, indicating that available potassium is the most representative in the direction of the available potassium-available phosphorus opposition variation represented by PC2. In the PC3 direction (contribution rate 15.32%), the absolute value of the loading of available phosphorus is the largest. After merging and deduplicating the representative indicators selected from each principal component, the core evaluation indicator set was finally determined to consist of four indicators: organic matter, available nitrogen, available potassium, and available phosphorus.

[0088] S103: Quantification of technical effect values ​​based on meta-analysis. This embodiment uses four indicators from the core evaluation indicator set—organic matter, alkaline nitrogen, available potassium, and available phosphorus—as the search targets. Statistical data of the experimental and control groups for each reduction technique on these four indicators are extracted from the literature dataset. The logarithmic response ratio and variance of each study record are calculated, and then input into a random effects model for weighted merging. The merged results are converted into percentage composite effect values, and the confidence intervals and significance results for each technique-indicator pair are determined.

[0089] Taking soil organic matter as an example, the meta-analysis results of various reduction technologies are shown below. Figure 4 . Figure 4 The overall effect and confidence interval of each reduction technique on soil organic matter content are presented in the form of a forest plot. The horizontal axis represents the contribution rate (%) of each technique to soil organic matter, and the vertical axis lists the names of each reduction technique. The diamond symbols in the plot represent the overall effect value of each group, the horizontal lines represent the corresponding confidence intervals, and the asterisks indicate the significance level (*** indicates p < 0.001, ** indicates p < 0.01). Figure 4 It can be seen that in the composite treatment category, the combined effect of organic fertilizer + multi-effect microbial fertilizer was 16.51%, with a confidence interval of [14.77%, 18.25%], the confidence interval did not contain zero and the interval length was less than 10%, indicating that this technology has a significant effect on increasing soil organic matter content and the estimation accuracy is high; the combined effect of earthworm castings + biochar + zeolite was 36.75%, with a confidence interval of [35.78%, 37.72%], the interval length was extremely short, and the conclusion had the highest reliability; the combined effect of organic fertilizer + compound fertilizer + urea was 2.81%, with a confidence interval of [1.65%, 3.97%], the effect was significant but the magnitude was relatively small. In the biological treatment category, the combined effect of earthworm castings also reached a significant level (p < 0.001). In the chemical treatment category, the combined effect of each soil conditioner treatment reached a significant level, but the effect magnitude was lower than that of the composite treatment category. The combined effect value and confidence interval of the above-mentioned technology-indicator pairs together constitute the data basis for the subsequent decision matrix.

[0090] S4: Constructing the TOPSIS evaluation model. In this embodiment, the comprehensive effect values ​​of each reduction technology on four indicators—organic matter, alkaline nitrogen, available potassium, and available phosphorus—are used as matrix elements. The reduction technology to be evaluated is represented by rows, and the four core indicators are represented by columns, to construct a basic decision matrix. In this embodiment, all four indicators are defined as positive indicators, and the basic decision matrix is ​​standardized using the extreme value standardization method to obtain a standardized matrix.

[0091] Subsequently, the standardized matrix was weighted using both the entropy weighting method and the analytic hierarchy process (AHP). The entropy weighting method calculates objective weights based on the degree of distributional difference among various indicators across different technologies, with indicators receiving higher objective weights for those exhibiting greater differences. The AHP method, specifically for the soil infertility barrier scenario in North China, involves domain experts setting subjective importance vectors for each indicator, constructing a consensus judgment matrix, and solving for the normalized eigenvectors to obtain subjective weights. The objective weights from the entropy weighting method and the subjective weights from the AHP method are linearly fused using a fusion coefficient α = 0.60 to obtain combined weights. The standardized matrix is ​​then multiplied by the combined weights to obtain the weighted decision matrix.

[0092] The computational principle of the TOPSIS decision model is described in [link to TOPSIS decision model]. Figure 5 . Figure 5 The core calculation logic of TOPSIS is presented in a two-dimensional diagram: the horizontal axis represents the increasing direction of benefit for indicator Y1, and the vertical axis represents the increasing direction of benefit for indicator Y2. Green dots represent positive ideal solutions Rj+ (hypothetical solutions where all indicators take optimal values), and red dots represent negative ideal solutions Rj- (hypothetical solutions where all indicators take worst values). Blue crosses T1, T2, and T3 represent the positions of the three reduction techniques to be evaluated in the two-dimensional indicator space. Di+ and Di- are the Euclidean distances between each technique and the positive and negative ideal solutions, respectively, and Ci is the formula for calculating relative proximity. Figure 5 It can be seen that T3 is closest to the positive ideal solution and furthest from the negative ideal solution. Therefore, T3 has the highest relative closeness and the best overall performance.

[0093] In introducing Monte Carlo uncertainty simulation, this invention uses the confidence intervals corresponding to each comprehensive effect value as the sampling range, randomly samples each matrix element in the basic decision matrix to generate a preset number of simulation decision matrices, and repeatedly standardizes and combines the weighted processing of each simulation decision matrix to obtain a corresponding number of weighted simulation decision matrices. Positive ideal solutions and negative ideal solutions are determined on each weighted simulation decision matrix, and the Euclidean distance and relative closeness between each reduction technique and the positive and negative ideal solutions are calculated. The mean of the relative closeness of each technique under all simulations is taken as the final relative closeness Ci.

[0094] S105: Suitability Ranking and Recommendation. Each reduction technique is ranked in descending order based on its final relative proximity score (Ci). A Ci=0.55 is used as the threshold separating priority recommendation from recommendation, and a Ci=0.45 is used as the threshold separating recommendation from optional. Techniques below 0.45 are further categorized as optional or not recommended. The final technique ranking results are shown below. Figure 6 The detailed evaluation and recommendation list is shown in Table 2.

[0095] Table 2 Final Evaluation and Recommendation Results of This Example

[0096] Depend on Figure 6 As can be seen, this embodiment sorted and graded 10 reduction techniques. The horizontal axis represents the suitability score Ci, and the vertical axis lists the names of each technique. The bar chart colors correspond to four levels: A (preferred, dark blue), B (recommended, medium blue), C (optional, light blue), and D (conditional / not recommended, gray). Figure 6 As shown in Table 3, the relative similarity Ci of organic fertilizer + multi-effect microbial fertilizer is 0.606, ranking first and rated as a Class A priority recommended technology. Its Euclidean distance D+ to the positive ideal solution is 0.4677, and its Euclidean distance D- to the negative ideal solution is 0.7203, indicating that this technology's comprehensive performance in the four core indicators of organic matter, alkaline nitrogen, available potassium, and available phosphorus is closest to the positive ideal solution. Earthworm castings + biochar + zeolite Ci=0.580, ranking second, also listed as a Class A priority recommended technology. Organic fertilizer + compound fertilizer + urea (Ci=0.543) and organic fertilizer + multi-effect microbial fertilizer + compound fertilizer are also recommended. Four technologies—urea + urea (Ci=0.532), vermicompost + biochar + zeolite + PAM (Ci=0.493), and liquid organic carbon + multi-effect microbial fertilizer (Ci=0.455)—are rated as Category B recommended. Three technologies—5 kg / 667 m² soil conditioner (Ci=0.392), chemical fertilizer + 15 t / hm² organic amendment (Ci=0.356), and application of vermicompost (Ci=0.352)—are rated as Category C optional. The combination of chemical fertilizer + 15 t / hm² organic amendment + 2.25 t / hm² inorganic amendment (Ci=0.342) is rated as Category D not recommended. Based on the above results, this invention recommends organic fertilizer + multi-effect microbial fertilizer as the most suitable soil reduction technology under the conditions of soil infertility in North China.

[0097] This invention employs a data-driven algorithm throughout, significantly reducing the interference of subjective judgment. By using PCA to screen indicators, meta-analysis to obtain effect values, and then ranking them using the TOPSIS model, the scientific rigor and reproducibility of the method are ensured. Furthermore, this invention simultaneously incorporates the effect values ​​and uncertainties obtained from meta-analysis into the TOPSIS decision-making framework, enabling decision-makers to consider both the benefits and risks of the technology. Through meta-analysis, this invention quantifies and integrates relevant publicly available data into local decision-making, ensuring that recommendations are based on solid scientific evidence. As new research literature is published, the meta-analysis database can be continuously updated, allowing for continuous optimization of evaluation results and enabling self-evolution. The method of this invention allows for large-scale, rapid paper simulations and prioritization before technology implementation, avoiding blindly conducting expensive field trials and saving time and economic costs.

[0098] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for evaluating the suitability of soil obstacle reduction technology in low- and medium-yield farmland, characterized in that, include: S1: Integrate the field sampling data and public data of the target area to obtain a preliminary index pool containing multiple soil obstacle-related indicators; S2: Input the initial index pool into the principal component analysis model to perform index dimensionality reduction and obtain the core evaluation index set; S3: Based on the random effects model, the effect data of multiple reduction techniques on multiple indicators in the core evaluation indicator set in the public data are analyzed to obtain the comprehensive effect value and corresponding confidence interval of multiple technical indicator pairs; S4: Construct a decision matrix using the comprehensive effect value as matrix elements, introduce combined weighting processing and Monte Carlo uncertainty simulation into the TOPSIS decision model, and calculate the relative closeness of multiple reduction techniques. S5: Soil obstacle reduction technologies are sorted in descending order based on their relative proximity to obtain suitability recommendations for soil obstacle reduction technologies.

2. The method for evaluating the suitability of soil obstacle reduction technology in low- and medium-yield farmland according to claim 1, characterized in that, Step S1 further includes: S11: Systematically set up sampling points in the target area, collect topsoil samples, and systematically measure the physical, chemical and biological properties of the soil to obtain field sampling data. S12: Based on the soil obstacle type in the target area, a systematic search of public databases is conducted according to the preset inclusion criteria to obtain a literature dataset; S13: Integrate the field sampling data with the literature dataset, and conduct preliminary screening of candidate indicators based on soil barrier mechanisms to obtain a preliminary indicator pool.

3. The method for evaluating the suitability of soil obstacle reduction technology in low- and medium-yield farmland according to claim 2, characterized in that, In step S12, the preset inclusion criteria include: Both experimental and control groups were set up. Core statistical information includes available or convertible sample size, mean, and standard deviation; The number of experiments repeated shall not be less than the preset number and the number of soil samples shall not be less than the preset quantity; The depth of the soil layer was uniformly limited to a preset depth range. The measurement units and standards for multiple indicators are consistent.

4. The method for evaluating the suitability of soil obstacle reduction technology in low- and medium-yield farmland according to claim 1, characterized in that, Step S2 further includes: S21: Standardize the sample data of the initial index pool and calculate the covariance matrix of the standardized data; S22: Perform eigenvalue decomposition on the covariance matrix, extract multiple principal components according to the cumulative variance contribution rate, and obtain the principal component loading matrix; S23: Analyze the principal component loading matrix, select the representative index with the largest absolute value of the loading from multiple principal components, merge the representative indices corresponding to multiple principal components to remove duplicates, and obtain the core evaluation index set.

5. The method for evaluating the suitability of soil obstacle reduction technology in low- and medium-yield farmland according to claim 1, characterized in that, Step S3 further includes: S31: Extract statistical data of the experimental and control groups of multiple indicators in the core evaluation index set from the publicly available dataset using multiple reduction techniques; S32: Based on the statistical data of the experimental and control groups for multiple indicators, the logarithm of the ratio of the experimental group mean to the control group mean is used as the logarithmic response ratio. The logarithmic response ratio and corresponding variance of multiple study records are calculated to obtain the effect size dataset of multiple study records. S33: Input the effect size dataset into the random effects model, merge multiple records of the same technical indicator pair, and convert the merged log response ratio into a percentage effect value to obtain the comprehensive effect value; S34: Determine the corresponding confidence interval based on the standard error corresponding to the comprehensive effect value, and determine the statistical significance of multiple technical indicators on the effect through significance testing.

6. The method for evaluating the suitability of soil obstacle reduction technology in low- and medium-yield farmland according to claim 1, characterized in that, Step S4 further includes: S41: Using the comprehensive effect value as matrix elements, with the reduction techniques to be evaluated as rows and multiple indicators from the core evaluation indicator set as columns, construct a basic decision matrix; S42: Standardize the basic decision matrix according to the attribute types of multiple indicators to obtain a standardized matrix; S43: The standardized matrix is ​​weighted using the entropy weighting method and the analytic hierarchy process, and the resulting objective weights and subjective weights are linearly fused to obtain the combined weights. S44: Multiply the standardized matrix by the combined weights to obtain the weighted decision matrix; S45: Using the confidence interval corresponding to the comprehensive effect value as the sampling range, perform Monte Carlo random sampling on the basic decision matrix to generate a preset number of simulated decision matrices, and perform combined weighting processing on multiple simulated decision matrices to obtain a corresponding number of weighted simulated decision matrices. S46: Determine the positive and negative ideal solutions for each index on multiple weighted simulation decision matrices, calculate the Euclidean distances between multiple reduction techniques and the positive and negative ideal solutions, and obtain the relative closeness of multiple reduction techniques.

7. The method for evaluating the suitability of soil obstacle reduction technology in low- and medium-yield farmland according to claim 1, characterized in that, In step S5, the recommendation levels of the suitability recommendation results include: preferred recommendation, recommended, optional, and not recommended.

8. A suitability evaluation system for soil obstacle reduction technology in low- and medium-yield farmland, characterized in that, include: The data acquisition module is used to integrate field sampling data and publicly available data from the target area to obtain a preliminary index pool containing multiple soil obstacle-related indicators. Dimensionality reduction module: used to input the initial index pool into the principal component analysis model, perform index dimensionality reduction, and obtain the core evaluation index set; Analysis module: used to analyze the effect data of multiple reduction techniques on multiple indicators in the core evaluation indicator set based on the random effects model, and to obtain the comprehensive effect value and corresponding confidence interval of multiple technical indicator pairs; The calculation module is used to construct a decision matrix with the combined effect value as matrix elements, introduce the combined weighting process and Monte Carlo uncertainty simulation into the TOPSIS decision model, and calculate the relative closeness of multiple reduction techniques. Output module: Used to sort multiple mitigation techniques in descending order based on their relative proximity, and obtain the suitability recommendation results for soil obstacle reduction techniques.

9. A suitability evaluation device for soil obstacle reduction technology in low- and medium-yield farmland, characterized in that, include: A memory and at least one processor, wherein the memory stores instructions; At least one of the processors invokes the instructions in the memory to cause a suitability evaluation device for soil obstacle reduction technology in low- and medium-yield fields to perform a suitability evaluation method for soil obstacle reduction technology in low- and medium-yield fields as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed by a processor, implement a suitability evaluation method for soil obstacle reduction technology in low- and medium-yield farmland as described in any one of claims 1 to 7.

Images