A farmland quality evaluation method and system based on adaptive weights

Abstract

The application discloses a kind of based on adaptive weight's cultivated land quality evaluation method and system, wherein, the method obtains the target cultivated land data of the cultivated land to be evaluated;Target cultivated land data is input into optimized cultivated land quality evaluation component to carry out quality evaluation, and the cultivated land quality evaluation result of cultivated land to be evaluated is obtained;Optimized cultivated land quality evaluation component is obtained by the following steps optimization: obtaining cultivated land training dataset;Cultivated land training dataset is input into tree model to carry out model training, and the trained tree model and the decision tree output by the trained tree model are obtained;According to cultivated land training dataset, all decision trees are extracted according to rule weight, and the target weight parameter is obtained;According to target weight parameter, cultivated land quality evaluation component is optimized and updated, and the optimized cultivated land quality evaluation component is obtained.The method can improve the explainability and accuracy of quality evaluation on cultivated land.The application relates to the technical field of soil quality evaluation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of soil quality evaluation, and in particular to a cultivated land quality evaluation method and system based on adaptive weights. BACKGROUND

[0002] With the rapid development of modern agriculture, cultivated land quality has become a key factor affecting agricultural production efficiency and sustainability. In order to fully understand and master the cultivated land quality situation and scientifically guide agricultural production, cultivated land quality grade evaluation has become one of the focuses of relevant personnel.

[0003] At present, the related technology is usually based on weight assignment methods such as expert scoring method and analytic hierarchy process to assign weights to various cultivated land elements, and evaluate the cultivated land quality based on the assigned weights. This method relies heavily on the subjective judgment of experts and is difficult to fully reflect the actual situation of different regions and different types of cultivated land. The explainability and accuracy of cultivated land quality evaluation are not satisfactory.

[0004] Therefore, the problems of the related technology need to be solved and optimized. SUMMARY

[0005] The purpose of the present application is to at least partially solve one of the technical problems in the related art.

[0006] To this end, one purpose of the embodiments of the present application is to provide a cultivated land quality evaluation method and system based on adaptive weights, which can improve the explainability and accuracy of quality evaluation of cultivated land.

[0007] In order to achieve the above technical purpose, the technical solutions adopted by the embodiments of the present application include:

[0008] In a first aspect, the embodiments of the present application provide a cultivated land quality evaluation method based on adaptive weights, comprising:

[0009] obtaining target cultivated land data of a cultivated land to be evaluated;

[0010] inputting the target cultivated land data into an optimized cultivated land quality evaluation component for quality evaluation to obtain a cultivated land quality evaluation result of the cultivated land to be evaluated;

[0011] wherein the optimized cultivated land quality evaluation component is obtained by the following steps:

[0012] obtaining a cultivated land training data set, the cultivated land training data set comprising a plurality of cultivated land data groups, each cultivated land data group comprising cultivated land element data and cultivated land quality quantification data;

[0013] inputting the cultivated land training data set into a tree model for model training to obtain a trained tree model and a plurality of decision trees output by the trained tree model, the decision trees being a tree representation set of a plurality of decision rules, the decision rules being used to indicate a nonlinear relationship between the cultivated land element data and the cultivated land quality quantification data;

[0014] extracting rule weights of all the decision trees according to the cultivated land training data set to obtain a plurality of target weight parameters, the target weight parameters being adaptive weight parameters of original linear features, the original linear features being element features of the cultivated land element data or rule features of the decision rules;

[0015] optimizing and updating original weight parameters in a cultivated land quality evaluation component according to all the target weight parameters to obtain the optimized cultivated land quality evaluation component.

[0016] In addition, the method according to the above-mentioned embodiments of the present application can have the following additional technical features:

[0017] Further, in an embodiment of the present application, the extracting rule weights of all the decision trees according to the cultivated land training data set to obtain a plurality of target weight parameters comprises:

[0018] obtaining element features of each of the cultivated land element data;

[0019] inputting the cultivated land training data set and all the decision trees into a rule model for rule feature extraction to obtain rule features output by the rule model;

[0020] performing weight linear analysis processing on all the element features and the rule features to obtain a plurality of the target weight parameters.

[0021] Further, in an embodiment of the present application, the inputting the cultivated land training data set and all the decision trees into a rule model for rule feature extraction to obtain rule features output by the rule model comprises:

[0022] extracting root leaf node paths of the decision trees through the rule model to obtain a plurality of decision paths;

[0023] performing rule feature processing on all the decision paths to obtain candidate rules corresponding to each of the decision paths;

[0024] encoding relationships of each of the cultivated land data groups in the cultivated land training data set according to all the candidate rules to obtain the rule features.

[0025] Further, in one embodiment of the present application, the weight linear analysis is performed on all the element features and the rule features to obtain a plurality of target weight parameters, including:

[0026] A feature matrix is constructed according to all the element features and the rule features;

[0027] A sparse linear regression model is constructed according to the cultivated land quality quantification data and the feature matrix;

[0028] Non-zero coefficients in the sparse linear regression model are extracted to obtain a plurality of intermediate weight parameters;

[0029] Importance optimization is performed on all the intermediate weight parameters to obtain a plurality of target weight parameters, each of which corresponds to an intermediate weight parameter.

[0030] Further, in one embodiment of the present application, the importance optimization is performed on all the intermediate weight parameters to obtain a plurality of target weight parameters, including:

[0031] Importance measurement is performed on all the intermediate weight parameters to obtain an importance measurement value of each intermediate weight parameter;

[0032] Normalization measurement is performed on all the original weight parameters in the cultivated land quality evaluation component to obtain a normalization measurement value of each original weight parameter;

[0033] According to the normalization measurement value and the importance measurement value, weight optimization is performed on the intermediate weight parameters to obtain the target weight parameters.

[0034] Further, in one embodiment of the present application, the weight optimization is performed on the intermediate weight parameters according to the normalization measurement value and the importance measurement value to obtain the target weight parameters, including:

[0035] According to the normalization measurement value, the importance measurement value is updated to obtain a target measurement value of the intermediate weight parameter;

[0036] According to the target measurement value, measurement optimization is performed on the intermediate weight parameters to obtain the target weight parameters.

[0037] Further, in one embodiment of the present application, the target cultivated land data is input into the optimized cultivated land quality evaluation component for quality evaluation to obtain a cultivated land quality evaluation result of the to-be-evaluated cultivated land, including:

[0038] A cultivated land quality grade system table is obtained;

[0039] inputting the target farmland data into the optimized farmland quality evaluation component to perform quality score calculation, to obtain a farmland quality score of the farmland to be evaluated;

[0040] performing grade division evaluation on the farmland quality score according to the farmland quality grade system table, to obtain the farmland quality evaluation result.

[0041] In a second aspect, an embodiment of the present application provides a farmland quality evaluation system based on adaptive weights, comprising:

[0042] a first processing unit configured to acquire target farmland data of a farmland to be evaluated;

[0043] a second processing unit configured to input the target farmland data into an optimized farmland quality evaluation component to perform quality evaluation, to obtain a farmland quality evaluation result of the farmland to be evaluated;

[0044] The optimized farmland quality evaluation component is obtained through the following steps:

[0045] acquiring a farmland training data set, the farmland training data set comprising a plurality of farmland data groups, each farmland data group comprising farmland element data and farmland quality quantization data;

[0046] inputting the farmland training data set into a tree model to perform model training, to obtain a trained tree model and a plurality of decision trees output by the trained tree model, the decision tree being a tree-shaped representation set of a plurality of decision rules, the decision rules being used to indicate a nonlinear relationship between the farmland element data and the farmland quality quantization data;

[0047] extracting rule weights from all the decision trees according to the farmland training data set, to obtain a plurality of target weight parameters, the target weight parameters being adaptive weight parameters of original linear features, the original linear features being element features of the farmland element data or rule features of the decision rules;

[0048] updating original weight parameters in a farmland quality evaluation component according to all the target weight parameters, to obtain the optimized farmland quality evaluation component.

[0049] In a third aspect, an embodiment of the present application further provides an electronic device, comprising:

[0050] at least one processor;

[0051] at least one memory configured to store at least one program;

[0052] When the at least one program is executed by the at least one processor, the at least one processor implements the method described above.

[0053] In a fourth aspect, the embodiments of the present application further provide a computer readable storage medium, wherein a processor executable program is stored, and the processor executable program is used for implementing the above method when executed by a processor.

[0054] The advantages and beneficial effects of the present application will be partially given in the following description, partially will become obvious from the following description, or will be learned by the practice of the present application:

[0055] The method disclosed by the embodiments of the present application comprises the following steps: obtaining target farmland data of a farmland to be evaluated; inputting the target farmland data into an optimized farmland quality evaluation component for quality evaluation to obtain a farmland quality evaluation result of the farmland to be evaluated; and obtaining a farmland training data set, wherein the farmland training data set comprises a plurality of farmland data groups, and each farmland data group comprises farmland element data and farmland quality quantification data; inputting the farmland training data set into a tree model for model training to obtain a trained tree model and a plurality of decision trees output by the trained tree model, wherein the decision tree is a tree representation set of a plurality of decision rules, and the decision rules are used to indicate a nonlinear relationship between the farmland element data and the farmland quality quantification data; extracting rule weights of all the decision trees according to the farmland training data set to obtain a plurality of target weight parameters, wherein the target weight parameters are adaptive weight parameters of original linear features, and the original linear features are element features of the farmland element data or rule features of the decision rules; and optimizing and updating original weight parameters in a farmland quality evaluation component according to all the target weight parameters to obtain the optimized farmland quality evaluation component. The method extracts weight parameters corresponding to each decision rule of the decision tree through a rule model, the decision rule records the nonlinear relationship between the farmland element data and the farmland quality quantification data, the extracted target weight parameters can indicate the contribution of the corresponding farmland element to the farmland quality, and then the original weight parameters in the farmland quality evaluation component are optimized and updated based on the adaptive target weight parameters, which is beneficial to improving the explainability and accuracy of the farmland quality evaluation and can more fully reflect the actual situation of different regions and different types. BRIEF DESCRIPTION OF DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following introduces the drawings of the related technical solutions in the embodiments of the present application or the prior art. It should be understood that the drawings in the following introduction are only for the convenience of expressing part of the embodiments of the technical solutions in the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the premise of the drawings.

[0057] Figure 1 A flowchart of a cultivated land quality evaluation method based on adaptive weight provided by the embodiments of the present application is shown in the figure.

[0058] Figure 2 An optimization flowchart of a cultivated land quality evaluation component provided by the embodiments of the present application is shown in the figure.

[0059] Figure 3 A framework diagram of a cultivated land quality evaluation system based on adaptive weight provided by the embodiments of the present application is shown in the figure.

[0060] Figure 4 A structural diagram of an electronic device provided by the embodiments of the present application is shown in the figure. DETAILED DESCRIPTION

[0061] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the drawings, wherein the same or similar reference signs represent the same or similar elements or elements with the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and are only used to explain the present application, and cannot be understood as limiting the present application. For the step numbers in the following embodiments, they are only set for the convenience of description and explanation, and the order between the steps is not limited in any way, and the execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0062] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application, and are not intended to limit the present application.

[0063] At present, the related technologies usually assign weights to each cultivated land element based on weight assignment methods such as expert scoring method and analytic hierarchy process, and evaluate the cultivated land quality based on the assigned weights, which seriously depends on the subjective judgment of experts and is difficult to fully reflect the actual situation of different regions and different types of cultivated land, and the interpretability and accuracy of the cultivated land quality evaluation are not satisfactory. In addition, the assigned weights determined by this method are usually fixed, which is difficult to adapt to the complex nonlinear relationship between cultivated land quality and cultivated land elements, and cannot fully reflect the dynamic changes of different cultivated land elements in different regions and cultivation projects, is not flexible enough, has limited scope of application, and has poor accuracy.

[0064] In addition, there are some related technologies that construct a machine learning model and directly predict the quality category of cultivated land data through the machine learning model to realize the quality evaluation of cultivated land. However, due to the black box effect of the machine learning model, the interpretability of the cultivated land quality evaluation result obtained by this method is poor. In order to ensure a high prediction accuracy of the machine learning model, it needs more training data, which has high training cost and long training time.

[0065] It should be noted that the above-mentioned related technologies are only used to assist in understanding the technical solutions of the present application and do not mean that they belong to the published prior art.

[0066] Therefore, the embodiments of the present application provide a cultivated land quality evaluation method and system based on adaptive weights, wherein the method extracts the weight parameters corresponding to each decision rule of the decision tree through a rule model, the decision rule records the nonlinear relationship between the cultivated land element data and the cultivated land quality quantization data, the target weight parameters extracted can indicate the contribution of the corresponding cultivated land element to the cultivated land quality, and then the original weight parameters in the cultivated land quality evaluation component are updated based on the target weight parameters, which is beneficial to improve the interpretability and accuracy of the cultivated land quality evaluation, can more fully reflect the actual situation of different regions and different types, has higher flexibility, and has wider scope of application.

[0067] In addition, the method learns the nonlinear relationship between the cultivated land elements and the cultivated land quality through a tree model and outputs the decision rules, and then extracts the weight parameters corresponding to each interpretable decision rule of the decision tree through a rule model, which can determine the contribution of each cultivated land element to the cultivated land quality in a data-driven manner, which not only reduces the training cost and the required training data, but also is beneficial to improve the interpretability and objectivity of the subsequent cultivated land quality evaluation.

[0068] Reference Figure 1 In the embodiments of the present application, a cultivated land quality evaluation method based on adaptive weights comprises:

[0069] Step 110, obtaining target farmland data of the farmland to be evaluated;

[0070] Step 120, inputting the target farmland data into the optimized farmland quality evaluation component for quality evaluation, to obtain a farmland quality evaluation result of the farmland to be evaluated;

[0071] In the embodiments of the present application, the farmland to be evaluated can be a paddy field, a watered field or a dry field to be evaluated, and the target farmland data can be a data set of various farmland elements of the farmland to be evaluated, which specifically includes stable farmland elements and variable farmland elements. The stable farmland elements include terrain slope, surface rock outcrop degree, depth of barrier layer from the ground surface and salinization degree. The variable farmland elements include field surface slope, groundwater level, effective soil layer thickness, profile configuration, irrigation guarantee rate, drainage condition, surface soil texture, organic matter content and soil acid-base degree. The farmland quality evaluation component can be a workflow execution tool constructed based on ModelBuilder technology, and the workflow execution tool records a plurality of weight parameters.

[0072] It can be understood that after obtaining the target farmland data of the farmland to be evaluated, the target farmland data can be input into the optimized farmland quality evaluation component, and the target farmland data is calculated based on the various weight parameters in the optimized farmland quality evaluation component. There are various specific calculation methods, such as weighted average calculation method, weighted summation calculation method, etc. The examples in the present application are for illustration only and do not limit the present application, so as to obtain the farmland quality evaluation result of the farmland to be evaluated.

[0073] In some embodiments, the inputting the target farmland data into the optimized farmland quality evaluation component for quality evaluation, to obtain the farmland quality evaluation result of the farmland to be evaluated, comprises:

[0074] obtaining a farmland quality grade system table;

[0075] inputting the target farmland data into the optimized farmland quality evaluation component for quality score calculation, to obtain a farmland quality score of the farmland to be evaluated;

[0076] performing grade division evaluation on the farmland quality score according to the farmland quality grade system table, to obtain the farmland quality evaluation result.

[0077] In the embodiments of the present application, the target farmland data can be input into the optimized farmland quality evaluation component for score calculation, to obtain the farmland quality score of the farmland to be evaluated. The farmland quality score specifically can include at least one of farmland crop quality score, farmland natural quality score, farmland utilization quality score and farmland economic quality score. Exemplarily, the equivalent expression of the farmland quality score specifically can be:

[0078]

[0079] wherein, is a cultivated land crop quality score, which is specifically a cultivated land natural quality score when the ith cultivated land to be evaluated plants the jth specified crop; is the total number of weight parameters; is the kth weight parameter in the optimized cultivated land quality evaluation component; is the total number of weight parameters in the optimized cultivated land quality evaluation component; is an index score of a cultivated land element corresponding to the kth weight parameter when the ith cultivated land to be evaluated plants the jth specified crop; is a cultivated land natural quality score of the ith cultivated land to be evaluated; is the total number of specified crops; is a production potential index of the jth specified crop; is a yield ratio coefficient of the jth specified crop; is a cultivated land utilization quality score of the ith cultivated land to be evaluated; is a cultivated land natural quality score when the ith cultivated land to be evaluated plants the jth specified crop; is a land utilization coefficient of the jth specified crop; is a cultivated land economic quality score of the ith cultivated land to be evaluated; is a land economic coefficient of the jth specified crop.

[0080] It can be understood that the production potential index , the yield ratio coefficient, the land utilization coefficient, the land economic coefficient and the like are known parameters, which can be determined by consulting existing information or obtained by using the calculatefield tool in the ArcGISToolbox toolbox to assign parameters in the process of modeling the cultivated land using the Modelbuilder modeling tool. There are various specific implementation manners, which will not be described here.

[0081] It should be noted that after the cultivated land quality score of the cultivated land to be evaluated is determined, the corresponding cultivated land quality level in the cultivated land quality level system table can be queried based on the cultivated land quality score. The cultivated land quality level can be at least one of a cultivated land natural quality level, a cultivated land utilization quality level, and a cultivated land economic quality level, so as to obtain the cultivated land quality evaluation result of the cultivated land to be evaluated.

[0082] Referring to Figure 2 , the optimized cultivated land quality evaluation component is obtained by the following steps:

[0083] Step 130, obtaining a cultivated land training data set, the cultivated land training data set comprising a plurality of cultivated land data groups, each cultivated land data group comprising cultivated land element data and cultivated land quality quantification data;

[0084] In the embodiments of the present application, the cultivated land training data set comprises a plurality of cultivated land data groups, each cultivated land data group corresponding to a cultivated land unit for training. Specifically, for a certain cultivated land data group, the cultivated land element data thereof can be a data set of various cultivated land elements of the corresponding cultivated land unit, and the specific cultivated land elements are similar to the cultivated land elements of the foregoing target cultivated land, which can be simply analogized; and the cultivated land quality quantification data can be a cultivated land quality score of the corresponding cultivated land unit, which is similar to the cultivated land quality score of the foregoing cultivated land to be evaluated, and the present application will not be described here.

[0085] Step 140, inputting the cultivated land training data set into a tree model for model training to obtain a trained tree model and a plurality of decision trees output by the trained tree model, the decision trees being a tree-shaped representation set of a plurality of decision rules, the decision rules being used to indicate a nonlinear relationship between the cultivated land element data and the cultivated land quality quantification data;

[0086] In the embodiments of the present application, the tree model can be a random forest (Random Forest) model, a gradient boosting tree (XGBoost) model, etc. Specifically, the cultivated land training data set can be input into the random forest model and the gradient boosting tree model respectively for training, and the nonlinear relationship between the cultivated land element data and the cultivated land quality quantification data is learned by the model, so as to obtain the trained random forest model and the trained gradient boosting tree model; then, the model with the highest prediction accuracy is selected as the trained tree model, and a plurality of decision trees output by the trained tree model are obtained.

[0087] Step 150, extracting rule weights of all the decision trees according to the cultivated land training data set to obtain a plurality of target weight parameters, the target weight parameters being adaptive weight parameters of original linear features, the original linear features being element features of the cultivated land element data or rule features of the decision rules;

[0088] In the embodiments of the present application, the nonlinear relationship recorded by all the decision trees can be encoded into a group of original linear features which are easy to be processed by a linear model based on the cultivated land training data set, and then the target weight parameters are determined based on the original linear features.

[0089] In some embodiments, the extracting rule weights of all the decision trees according to the cultivated land training data set to obtain a plurality of target weight parameters comprises:

[0090] obtaining an element feature of each of the cultivated land element data;

[0091] In the embodiment of the present application, after preprocessing each cultivated land element data in the cultivated land training data set, the element feature of each cultivated land element data is obtained by feature extraction. There are various ways of feature extraction, which will not be described here.

[0092] The cultivated land training data set and all the decision trees are input into a rule model for rule feature extraction to obtain rule features output by the rule model;

[0093] Further, the input of the cultivated land training data set and all the decision trees into the rule model for rule feature extraction to obtain rule features output by the rule model comprises:

[0094] The rule model extracts the root leaf node path of the decision tree to obtain a plurality of decision paths;

[0095] All the decision paths are subjected to rule feature processing to obtain a candidate rule corresponding to each of the decision paths;

[0096] According to all the candidate rules, each cultivated land data group in the cultivated land training data set is subjected to relationship coding to obtain the rule features.

[0097] In the embodiment of the present application, the rule model can be a RuleFit model. Specifically, for any decision tree, the rule model can extract all paths from the root node to each leaf node of the decision tree, which is denoted as a decision path. Each decision path represents a decision rule of the decision tree record, which includes a plurality of features, feature thresholds, and logical conditions, such as "decision rule R1: IF effective soil layer thickness > 100 cm AND organic matter content > 20 g / kg".

[0098] It can be understood that the rule feature processing can be to featureize the decision rule corresponding to each decision path to obtain a candidate rule corresponding to each decision path. For any candidate rule, the relationship coding can be to calculate the rule feature value between each cultivated land data group and the candidate rule, and then to code each rule feature value based on the candidate rule to obtain a plurality of rule features corresponding to the candidate rule. Each rule feature records the candidate rule and the corresponding rule feature value. The remaining candidate rules can be similarly deduced. Specifically, if a cultivated land data group satisfies all conditions in the candidate rule, the rule feature value between the cultivated land data group and the candidate rule is 1, otherwise, it is 0.

[0099] The weight linear analysis processing is performed on all the element features and the rule features to obtain a plurality of target weight parameters.

[0100] Further, the weight linear analysis processing on all the element features and the rule features to obtain a plurality of target weight parameters comprises:

[0101] A feature matrix is constructed according to all the element features and the rule features.

[0102] A sparse linear regression model is constructed according to the cultivated land quality quantification data and the feature matrix.

[0103] Non-zero coefficients in the sparse linear regression model are extracted to obtain a plurality of intermediate weight parameters.

[0104] In the embodiments of the present application, after obtaining all the element features and the rule features, the element features and the rule features can be merged to form an expanded feature matrix. The matrix columns of the feature matrix include the element features and the rule features, and the matrix rows of the feature matrix correspond to each cultivated land data group.

[0105] It can be understood that the cultivated land quality quantification data can be taken as a target variable, and the feature matrix can be taken as an input prediction variable to fit and construct a sparse linear regression model. The sparse linear regression model can be obtained based on Lasso (Least Absolute Shrinkage and Selection Operator) regression. After the sparse linear regression model is constructed, a plurality of non-zero coefficients can be extracted from the model parameters of the sparse linear regression model, and each non-zero coefficient is recorded as an intermediate weight parameter. Each intermediate weight parameter corresponds to an original linear feature selected by the sparse linear regression model, and the intermediate weight parameter is used to indicate the importance of the corresponding original linear feature.

[0106] Importance optimization processing is performed on all the intermediate weight parameters to obtain a plurality of target weight parameters. Each target weight parameter corresponds to an intermediate weight parameter.

[0107] Further, the importance optimization processing on all the intermediate weight parameters to obtain a plurality of target weight parameters comprises:

[0108] Importance measurement is performed on all the intermediate weight parameters to obtain an importance measurement value of each intermediate weight parameter.

[0109] Normalization measurement is performed on all the original weight parameters in the cultivated land quality evaluation component to obtain a normalization measurement value of each original weight parameter.

[0110] In the embodiments of the present application, the importance measure can first be to obtain the absolute value of each intermediate weight parameter, and then normalize the absolute values of all intermediate weight parameters, such as Min-Max normalization, to obtain the importance measure value corresponding to each intermediate weight parameter, which is used to represent the relative importance of the corresponding intermediate weight parameter in all intermediate weight parameters.

[0111] It can be understood that the normalization measure can first be to obtain all weight parameters used in the cultivated land quality evaluation component, denoted as original weight parameters, and then normalize all original weight parameters to obtain the normalization measure value of each original weight parameter, and the sum of all normalization measure values is 1.

[0112] According to the normalization measure value and the importance measure value, the intermediate weight parameter is weight-optimized to obtain the target weight parameter.

[0113] Further, the weight optimization of the intermediate weight parameter according to the normalization measure value and the importance measure value to obtain the target weight parameter comprises:

[0114] According to the normalization measure value, the importance measure value is updated to obtain the target measure value of the intermediate weight parameter;

[0115] According to the target measure value, the intermediate weight parameter is measure-optimized to obtain the target weight parameter.

[0116] In the embodiments of the present application, for any intermediate weight parameter, in a first implementation, the importance measure value can be updated based on the normalization measure value, such as calculating the normalization measure value and the importance measure value in a weighted average manner, and the weight used in the specific calculation process can be set according to actual conditions, to obtain the target measure value of the intermediate weight parameter; and then the intermediate weight parameter is measure-optimized through the target measure value, such as multiplying the target measure value and the intermediate weight parameter to obtain the target weight parameter.

[0117] Alternatively, in a second implementation, after determining the target measure value, the measure optimization of the intermediate weight parameter can be to replace the intermediate weight parameter with the target measure value, and determine the target measure value as the target weight parameter.

[0118] Step 160, according to all the target weight parameters, the original weight parameters in the cultivated land quality evaluation component are updated and optimized to obtain the optimized cultivated land quality evaluation component.

[0119] In the embodiment of the present application, for any one target weight parameter, the corresponding original weight parameter corresponding to the target weight parameter can be determined first, and the target weight parameter and the corresponding original weight parameter are both used for the same cultivated land element. Then, the corresponding original weight parameter in the cultivated land quality evaluation component is replaced using the target weight parameter, and the remaining target weight parameters are the same. In this way, the original weight parameters in the cultivated land quality evaluation component are updated and optimized to obtain an optimized cultivated land quality evaluation component.

[0120] A cultivated land quality evaluation system based on adaptive weight is described in detail below with reference to the accompanying drawings according to the embodiment of the present application.

[0121] Referring to Figure 3 The cultivated land quality evaluation system based on adaptive weight proposed in the embodiment of the present application comprises:

[0122] The first processing unit 101 is configured to obtain target cultivated land data of a cultivated land to be evaluated.

[0123] The second processing unit 102 is configured to input the target cultivated land data into the optimized cultivated land quality evaluation component for quality evaluation to obtain a cultivated land quality evaluation result of the cultivated land to be evaluated.

[0124] The optimized cultivated land quality evaluation component is obtained by the following steps:

[0125] The cultivated land training data set is obtained, and the cultivated land training data set comprises a plurality of cultivated land data groups, and each cultivated land data group comprises cultivated land element data and cultivated land quality quantization data.

[0126] The cultivated land training data set is input into a tree model for model training to obtain a trained tree model and a plurality of decision trees output by the trained tree model. The decision tree is a tree representation set of a plurality of decision rules, and the decision rule is used to indicate a nonlinear relationship between the cultivated land element data and the cultivated land quality quantization data.

[0127] According to the cultivated land training data set, rule weight extraction is performed on all the decision trees to obtain a plurality of target weight parameters. The target weight parameter is an adaptive weight parameter of an original linear feature. The original linear feature is an element feature of the cultivated land element data or a rule feature of the decision rule.

[0128] According to all the target weight parameters, the original weight parameters in the cultivated land quality evaluation component are updated and optimized to obtain the optimized cultivated land quality evaluation component.

[0129] It can be understood that the contents in the above method embodiments are applicable to the present system embodiments, the present system embodiments specifically implement the functions same as those of the above method embodiments, and achieve the beneficial effects same as those of the above method embodiments.

[0130] With reference to Figure 4 The present application embodiment further provides an electronic device, comprising:

[0131] at least one processor 201;

[0132] at least one memory 202, configured to store at least one program;

[0133] When the at least one program is executed by the at least one processor 201, the at least one processor 201 implements the above method embodiments.

[0134] Similarly, it can be understood that the contents in the above method embodiments are applicable to the present device embodiments, the present device embodiments specifically implement the functions same as those of the above method embodiments, and achieve the beneficial effects same as those of the above method embodiments.

[0135] The present application embodiment further provides a computer readable storage medium, wherein the computer readable storage medium stores a program executable by the processor 201, and the program executable by the processor 201, when executed by the processor 201, is used for implementing the above method embodiments.

[0136] Similarly, the contents in the above method embodiments are applicable to the present computer readable storage medium embodiments, the present computer readable storage medium embodiments specifically implement the functions same as those of the above method embodiments, and achieve the beneficial effects same as those of the above method embodiments.

[0137] In some alternative embodiments, the functions / operations mentioned in the block diagram can not occur in the order mentioned in the operation diagram. For example, depending on the functions / operations involved, two blocks shown in succession can actually be executed substantially simultaneously or the blocks can sometimes be executed in reverse order. In addition, the embodiments presented and described in the flowcharts of the present application are provided by way of example, with the purpose of providing a more comprehensive understanding of the technology. The disclosed method is not limited to the operations and logical flows presented herein. Alternative embodiments are contemplated in which the order of various operations is changed and in which sub-operations described as part of larger operations are independently executed.

[0138] Furthermore, although the present application is described in the context of functional modules, it is understood that one or more of the functions and / or features can be integrated in a single physical device and / or software module, or one or more functions and / or features can be implemented in separate physical devices or software modules. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary to an understanding of the present application. Rather, the actual implementation is within the routine skill of engineers familiar with the property, function and internal relationships of the various functional modules disclosed herein. Accordingly, the present application is not limited to the specific details of the functional modules described herein. Rather, it is understood that the skilled artisan, using ordinary skill in the art, can implement the application as taught in the claims without undue experimentation. It is also understood that the specific concepts disclosed are merely illustrative and that the scope of the present application is determined by the appended claims and their equivalents.

[0139] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part of the prior art that contributes to the technical solutions or the part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method of the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0140] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered a list of executable instructions for implementing logic functions, and can be specifically embodied in any computer-readable medium for use by an instruction execution system, apparatus, or device, such as a computer-based system, a system including a processor, or other system that can take instructions from an instruction execution system, apparatus, or device and execute them, or in conjunction with these instruction execution systems, apparatuses, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

[0141] More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection (electronic) having one or more wires, a portable computer diskette (magnetic), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can also be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, for example, via optical scanning of the paper or other medium, then compiled, interpreted, or otherwise processed in a suitable manner, if necessary, and then stored in a computer memory.

[0142] It should be understood that aspects of the application can be implemented in hardware, software, firmware or combinations thereof. In the above embodiments, various steps or methods can be implemented in software or firmware that is stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any of the following technologies, or combinations thereof, can be used: a discrete logic circuit having logic gates for implementing logic functions upon data signals, an application specific integrated circuit having appropriate combinational logic gates, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0143] In the above description of the present specification, the description referring to the terms "one embodiment", "another embodiment", or "certain embodiments" or the like means that a specific feature, structure, material or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present specification. The illustrative expressions of the above terms do not necessarily refer to the same embodiment or example throughout the present specification. Also, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0144] Although the embodiments of the present application have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, alternatives and variations to these embodiments can be made without departing from the principles and spirit of the application, the scope of which is defined by the claims and their equivalents.

[0145] The above is a specific description of the preferred embodiments of the present application, but the present application is not limited to the embodiments, and those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present application, and these equivalent modifications or substitutions are included in the scope defined by the claims of the present application.

Claims

1. A method for evaluating the quality of cultivated land based on adaptive weights, characterized by, The method comprises the following steps: obtaining target farmland data of a farmland to be evaluated; inputting the target farmland data into an optimized farmland quality evaluation component for quality evaluation to obtain a farmland quality evaluation result of the farmland to be evaluated; wherein the optimized farmland quality evaluation component is obtained through the following steps: obtaining a farmland training data set, the farmland training data set comprising a plurality of farmland data groups, each farmland data group comprising farmland element data and farmland quality quantification data; inputting the farmland training data set into a tree model for model training to obtain a trained tree model, and a plurality of decision trees output by the trained tree model, the decision tree being a tree representation set of a plurality of decision rules, the decision rules being used to indicate a nonlinear relationship between the farmland element data and the farmland quality quantification data; extracting rule weights from all the decision trees according to the farmland training data set to obtain a plurality of target weight parameters, the target weight parameters being adaptive weight parameters of original linear features, the original linear features being element features of the farmland element data or rule features of the decision rules; optimizing and updating original weight parameters in a farmland quality evaluation component according to all the target weight parameters to obtain the optimized farmland quality evaluation component; the extracting rule weights from all the decision trees according to the farmland training data set to obtain a plurality of target weight parameters comprises: obtaining element features of each farmland element data; inputting the farmland training data set and all the decision trees into a rule model for rule feature extraction to obtain rule features output by the rule model; performing weight linear analysis processing on all the element features and the rule features to obtain a plurality of the target weight parameters; the inputting the farmland training data set and all the decision trees into a rule model for rule feature extraction to obtain rule features output by the rule model comprises: extracting root leaf node paths of the decision trees through the rule model to obtain a plurality of decision paths; the rule model is a RuleFit model; performing rule feature processing on all the decision paths to obtain candidate rules corresponding to each decision path; according to all the candidate rules, performing relationship encoding on each farmland data group in the farmland training data set, calculating rule feature values between each farmland data group and the candidate rules through the relationship encoding, and respectively encoding each rule feature value based on the candidate rules to obtain the rule features.

2. The method of claim 1, wherein, the performing weight linear analysis processing on all the element features and the rule features to obtain a plurality of the target weight parameters comprises: constructing a feature matrix according to all the element features and the rule features; constructing a sparse linear regression model according to the farmland quality quantification data and the feature matrix; extracting non-zero coefficients in the sparse linear regression model to obtain a plurality of intermediate weight parameters; The importance optimization processing is performed on all the intermediate weight parameters to obtain a plurality of target weight parameters, each of which corresponds to one of the intermediate weight parameters.

3. The method of claim 2, wherein, The importance optimization processing is performed on all the intermediate weight parameters to obtain a plurality of target weight parameters, each of which corresponds to one of the intermediate weight parameters. The importance of all the intermediate weight parameters is measured to obtain an importance measurement value of each intermediate weight parameter. The normalization measurement is performed on all the original weight parameters in the cultivated land quality evaluation component to obtain a normalization measurement value of each original weight parameter. The weight optimization is performed on the intermediate weight parameters according to the normalization measurement value and the importance measurement value to obtain the target weight parameters.

4. The method of claim 3, wherein, The weight optimization is performed on the intermediate weight parameters according to the normalization measurement value and the importance measurement value to obtain the target weight parameters. The target measurement value of the intermediate weight parameter is obtained by updating the importance measurement value according to the normalization measurement value. The target weight parameters are obtained by performing measurement optimization on the intermediate weight parameters according to the target measurement value.

5. The method of claim 1, wherein, The target cultivated land quality evaluation result of the cultivated land to be evaluated is obtained by inputting the target cultivated land data into the optimized cultivated land quality evaluation component for quality evaluation, including: obtaining a cultivated land quality grade system table; calculating a quality score of the cultivated land to be evaluated by inputting the target cultivated land data into the optimized cultivated land quality evaluation component; grading and evaluating the quality score of the cultivated land according to the cultivated land quality grade system table to obtain the cultivated land quality evaluation result.

6. An adaptive weight-based cultivated land quality evaluation system characterized by comprising: including: a first processing unit configured to obtain target cultivated land data of cultivated land to be evaluated; a second processing unit configured to input the target cultivated land data into an optimized cultivated land quality evaluation component for quality evaluation to obtain a cultivated land quality evaluation result of the cultivated land to be evaluated; wherein the optimized cultivated land quality evaluation component is obtained by the following steps: obtaining a cultivated land training data set, the cultivated land training data set including a plurality of cultivated land data groups, each of which including cultivated land element data and cultivated land quality quantization data; inputting the cultivated land training data set into a tree model for model training to obtain a trained tree model and a plurality of decision trees output by the trained tree model, the decision tree being a tree representation set of a plurality of decision rules, the decision rules being used to indicate a nonlinear relationship between the cultivated land element data and the cultivated land quality quantization data; extracting rule weights from all the decision trees according to the cultivated land training data set to obtain a plurality of target weight parameters, the target weight parameters being self-adaptive weight parameters of original linear features, the original linear features being element features of the cultivated land element data or rule features of the decision rules; updating original weight parameters in a cultivated land quality evaluation component according to all the target weight parameters to obtain the optimized cultivated land quality evaluation component; and inputting the target cultivated land data into the optimized cultivated land quality evaluation component for quality evaluation to obtain the cultivated land quality evaluation result of the cultivated land to be evaluated. The rule weight extraction is performed on all the decision trees according to the cultivated land training data set, and a plurality of target weight parameters are obtained, including: obtaining the element characteristics of each cultivated land element data; inputting the cultivated land training data set and all the decision trees into a rule model for rule feature extraction, and obtaining rule features output by the rule model; performing weight linear analysis processing on all the element characteristics and the rule features, and obtaining a plurality of target weight parameters; The rule weight extraction is performed on all the decision trees according to the cultivated land training data set, and a plurality of target weight parameters are obtained, including: extracting the root leaf node path of the decision tree through the rule model, and obtaining a plurality of decision paths; the rule model is a RuleFit model; performing rule feature processing on all the decision paths, and obtaining candidate rules corresponding to each decision path; According to all the candidate rules, each cultivated land data group in the cultivated land training data set is relationally encoded, the rule feature value between each cultivated land data group and the candidate rules is calculated through the relational encoding, and each rule feature value is encoded based on the candidate rules, and the rule features are obtained.

7. An electronic device, comprising: including: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method of any one of claims 1-5.

8. A computer readable storage medium having stored therein a program which is executable by a processor, characterized in that, The program executable by the processor is used to implement the method of any one of claims 1-5 when executed by the processor.

Images