Intrusion regulation and control method for evaluating foreign fish invasion mechanism based on SHAP

Abstract

The invention belongs to the technical field of ecological environment monitoring and biological invasion risk assessment, and discloses an invasion regulation and control method for assessing an alien fish invasion mechanism based on SHAP, and the method comprises the following steps: (1) monitoring the type data of biocenosis in a target river network region and the biological characteristic data of each biocenosis; (2) selecting a proper model to construct an intrusion intensity prediction model; and (3) combining the intrusion intensity prediction model with an SHAP analysis algorithm to establish an intrusion analysis model, analyzing an intrusion response type and an intrusion response priority of the local water biocenosis invading the fishes to the outside through the intrusion analysis model, and generating a corresponding ecological regulation and control strategy according to the intrusion response type. Marking implementation priorities of the ecological regulation and control strategies according to the intrusion response priorities; wherein the intrusion response type comprises a promotion response and an inhibition response; the ecological regulation and control strategy comprises a weakening regulation and control strategy and a strengthening regulation and control strategy.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of ecological environment monitoring and biological invasion risk assessment, and particularly relates to an invasion control method based on SHAP evaluation of invasion mechanism of alien fish. BACKGROUND

[0002] Freshwater ecosystems are one of the most concentrated ecological units of global biodiversity, and are also the most significant ecological type affected by human activities. With the intensification of urbanization, river channelization, agricultural runoff and the input of alien species, the structure and function of river ecosystems are continuously degrading, and alien invasive fish has become an important factor threatening the stability of native communities and ecological security. Invasive fish often changes the original energy flow and trophic level structure through predation competition, niche replacement and food web restructuring, causing ecosystem homogenization and function loss. In urban rivers, the invasion process is often superimposed with multiple human disturbances (such as hydrological regulation, pollution load and heat discharge), which weakens the resistance of native species, allowing broad-adapted and tolerant alien species to quickly dominate, forming a typical "multiple stress coupling" effect. The invasion mechanism under such complex disturbances is complex and variable, and traditional single-factor research is difficult to fully reveal its ecological process.

[0003] Traditional ecological invasion research relies on multiple regression or single-factor statistical methods, which can describe local relationships but are difficult to handle high-dimensional, nonlinear, and strongly interactive ecological data. In recent years, the application of machine learning in ecology has provided new means for complex system modeling and has been widely used in species distribution prediction, ecosystem state assessment, water quality and invasion risk analysis. However, these models, although with high accuracy, have the "black box" problem, making it difficult to explain the causal relationship between feature variables and ecological responses, limiting their application in ecological decision-making and river management. Therefore, explainable artificial intelligence (XAI) technology has emerged, among which the Shapley additive explanation algorithm (SHAP) based on cooperative game theory has theoretical advantages such as consistency, additivity and local accuracy, which can quantify the marginal contribution of each feature to model prediction, thus revealing the key ecological driving factors, response direction and threshold effect. SHAP has been verified in large invertebrate distribution, seagrass habitat, estuarine water quality and soil organic carbon prediction, providing a reliable way for mechanism explanation and management application of ecological models.

[0004] In the multi-interference ecosystem such as urban river, it is difficult to depict the complex influence process of alien fish invasion by relying on single group or single index. In recent years, the ecological informatics research emphasizes the integration of "multi-community index" and "model interpretability". By combining traditional electric capture investigation and environmental DNA macro-barcode technology (eDNA), community structure data across trophic levels can be obtained. Through interpretable machine learning model, the invasion intensity quantification and multi-trophic level response mechanism analysis can be realized at the same time. This method helps to reveal the action pathway of alien invasive fish on different trophic levels (such as plankton, benthic invertebrates, algae, aquatic plants, etc.), and convert the mechanism results into operational invasion control and ecological restoration strategies. In summary, although existing research has made progress in ecological invasion prediction, there is still a lack of a systematic analysis framework with high-dimensional modeling ability and ecological interpretability. In this context, the present invention is proposed to build an analysis method for river alien invasive fish intensity and its influence mechanism on multi-trophic level community based on SHAP interpretability algorithm, realizing the integration of multi-source community data to mechanism identification and management decision support. SUMMARY

[0005] In view of this, in order to solve the problems proposed in the background art, the purpose of the present invention is to provide an invasion control method based on SHAP evaluation of alien fish invasion mechanism.

[0006] To achieve the above purpose, the present invention provides the following technical solutions.

[0007] An invasion control method based on SHAP evaluation of alien fish invasion mechanism, comprising the following steps:

[0008] (1) Data monitoring; monitoring the biological community species data and the biological characteristic data of each biological community in the target river network area;

[0009] (2) Selecting a suitable model to build an invasion intensity prediction model;

[0010] Inputting the data monitored in step (1) into the alternative model, and calculating the accuracy Accuracy, F1 value, AUC value and comprehensive evaluation function value of each alternative model Selecting a suitable model;

[0011] Evaluation function: ; In the formula, Macro-average area under the receiver operating characteristic curve; Macro-average F1 score;

[0012] (3) combining the invasion intensity prediction model with a SHAP analysis algorithm to establish an invasion analysis model, analyzing the invasion response type and the invasion response priority of the native aquatic biota community to the invasive fish through the invasion analysis model, generating a corresponding ecological regulation strategy according to the invasion response type, and marking the implementation priority of the ecological regulation strategy according to the invasion response priority;

[0013] wherein:

[0014] The invasion response type includes a promotion response and an inhibition response.

[0015] The ecological regulation strategy includes a weakening regulation strategy generated according to the promotion response and a strengthening regulation strategy generated according to the inhibition response.

[0016] Preferably, the biota species data includes fish, plankton, phytoplankton, macrobenthic invertebrates, aquatic higher plants and periphytic algae; and the biological characteristic data includes density , biomass , species number , Shannon diversity , Pielou evenness and Margalef richness .

[0017] .

[0018] Preferably, the data monitored in step (1) is further subjected to data preprocessing before being input into the alternative model:

[0019] The biological characteristic data not detected is subjected to zero processing;

[0020] The detected biological characteristic data is subjected to logarithmic transformation and robust scaling processing;

[0021] ; wherein, is the original value after logarithmic transformation; is the standardized value after robust scaling; is the median of the total variable population; is the interquartile range of the variable.

[0022] Preferably, the accuracy Accuracy, F1 value and AUC value in step (2) are calculated according to the following formula:

[0023] ;

[0024] wherein, TP is the number of true positives; TN is the number of true negatives; FP is the number of false positives; FN is the number of false negatives; TPR is the true positive rate; and FPR is the false positive rate.

[0025] Preferably, the invasion analysis model is expressed as:

[0026] , ; wherein: is the prediction result output by the invasion intensity prediction model , is the marginal contribution of the feature to the prediction result, is the total number of features, global importance is used to measure the average contribution of the feature to the overall prediction result, intra-class contribution is used to measure the prediction contribution of the feature to the prediction result of different classes.

[0027] Preferably, the step of analyzing the invasion response type of the native aquatic biota community to the alien invasive fish by the invasion analysis model comprises:

[0028] calculating the weighted average contribution of the target feature under all samples ;

[0029] , the invasion response type of the target feature is promotion response; , the invasion response type of the target feature is inhibition response;

[0030] wherein, ; wherein, is the weight of the prediction result of different classes.

[0031] Preferably, the invasion response type comprises:

[0032] independent promotion response and independent inhibition response for a single feature;

[0033] joint promotion response and joint inhibition response for at least two features.

[0034] Preferably, the ecological regulation strategy comprises:

[0035] independent weakening regulation strategy generated according to the independent promotion response and independent strengthening regulation strategy generated according to the independent inhibition response;

[0036] joint weakening regulation strategy generated according to the joint promotion response and joint strengthening regulation strategy generated according to the joint inhibition response.

[0037] Preferably, the step of analyzing the invasion response priority of the native aquatic biota community to the alien invasive fish by the invasion analysis model comprises:

[0038] calculating a community priority score and determining an invasion response priority according to the priority score ; wherein, ; wherein, is an adjustment coefficient, and the value range of the adjustment coefficient is 0.5-1.

[0039] Preferably, the adjustment coefficient is calculated by the following formula:

[0040] ; wherein, is a feature is the variance of the contribution degree within the class under the prediction result of different categories, is the absolute value of the average contribution of the feature population, is a smoothing constant.

[0041] Compared with the prior art, the present application has the following beneficial effects:

[0042] The present application provides an invasion control method based on SHAP evaluation of the invasion mechanism of alien fish, which analyzes and evaluates the influence mechanism of alien invasive fish on native aquatic biological community in rivers through SHAP explainable algorithm, and maps the threshold / contribution to control classification and management priority, realizes the quantitative execution from algorithm explanation to ecological control; through identifying the response direction and interaction of different trophic level communities to alien invasive fish, it provides quantitative basis for river restoration and invasion prevention and control; an automatically updated community level management system is established, which improves the scientificity and adaptability of river ecosystem management; effectively avoids subjective bias caused by traditional experience judgment and single factor decision, and realizes data-driven precise ecological regulation. BRIEF DESCRIPTION OF DRAWINGS

[0043] Figure 1 is a flow chart of the invasion control method based on SHAP evaluation of the invasion mechanism of alien fish according to the present application. DETAILED DESCRIPTION

[0044] For further understanding of the present application, the present application is described in detail in conjunction with the drawings and examples. The structures, proportions, sizes and the like shown in the drawings of the present specification are only used to cooperate with the content disclosed in the specification, to be understood and read by those skilled in the art, and are not used to limit the limiting conditions of the implementation of the present application, so they do not have technical significance. Any modification of structure, change of proportion relationship or adjustment of size, without affecting the effect and purpose that can be achieved by the present application, should still fall within the scope of the technical content disclosed by the present application. At the same time, the terms such as "upper", "lower", "left", "right", "middle" and the like used in the specification are only for the convenience of clear description, and are not used to limit the scope of implementation, and the change or adjustment of the relative relationship is also considered as the implementation of the present application without substantially changing the technical content. It should be noted that the terms "first", "second" and the like in the specification and claims of the present application and the above drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein.

[0045] As shown in Figure 1 , the present application provides an invasion control method based on SHAP evaluation of the invasion mechanism of alien fish, comprising the following steps:

[0046] (1) Data monitoring; monitoring the biological community species data and the biological characteristic data of each biological community in the target river network area;

[0047] Laying several fixed stations (denoted as sample set ) in the target river network area, it is suggested to repeat the monitoring according to the frequency of rainy / dry season, so as to capture the seasonal difference and improve the statistical efficiency.

[0048] The biological community species data monitored covers six aquatic biological communities: fish (F), zooplankton (Z), phytoplankton (P), large benthic invertebrates (BM), aquatic higher plants (AP), and periphytic algae (PA); the biological characteristic data of each biological community includes density , biomass , species number , Shannon diversity , Pielou evenness , and Margalef richness ; a unified feature space of 36 ecological indicators is constructed, and the biological characteristic data is specifically expressed as:

[0049] ; wherein, is the species in the community The relative abundance of each species.

[0050] To eliminate dimensional differences and extreme value effects, all continuous variables are logarithmically transformed and robustly scaled: ; In the formula, is the original value after logarithmic transformation; is the standardized value after robust scaling; is the median of the total variable; is the interquartile range of the variable (i.e. the difference between the 75th percentile and the 25th percentile).

[0051] To represent the fact of "community absence" rather than statistical absence, the true undetected biological characteristic data is set to zero, thus maintaining the 36-dimensional complete feature input and avoiding the loss of potential ecological signals due to multicollinearity (improving model robustness).

[0052] In addition, fish are sampled using a backpack electric trap and eDNA macro-barcode in parallel (morphology + molecular double evidence); the rest of the community is executed according to the current ecological monitoring specifications. The original eDNA sequence is obtained by conventional quality control, splicing and clustering process to obtain the target OTU / read count; non-target groups are excluded, and only fish sequences are retained for subsequent invasion judgment.

[0053] Relative abundance of electric capture (RAE) and relative abundance of eDNA (RAeD):

[0054] wherein, and are the number of individuals of alien invasive species and the total number of individuals, respectively; and are the environmental DNA sequence abundance counts of alien invasive fish and all fish. The invasion gradient label based on RAE and RAeD is classified according to the following gradient: high: RAE> 50% and RAeD> 50%; medium: RAE> 50% or RAeD> 50%; low: both RAE< 50% and RAeD< 50%.

[0055] (2) Select a suitable model to build an invasion intensity prediction model;

[0056] The present application adopts four types of classification models based on ensemble tree algorithm as alternative models, which are: random forest (Random Forest, RF), extreme gradient boosting (XGBoost), light gradient boosting (LightGBM) and category boosting (CatBoost).

[0057] The data monitored in step (1) is divided into a training set (80%) and a test set (20%); the training set is used to train each alternative model, and the test set is used to test the trained alternative model.

[0058] Model training: The training set data (36-dimensional feature matrix) is input into the alternative model, and the invasion gradient category of the alien invasive fish in the river is predicted by the alternative model For the predicted output results and the relative abundance of electric capture and eDNA, the training model parameters are optimized

[0059] Model testing and selection: The test set data (36-dimensional feature matrix) is input into the alternative model, and the invasion gradient category of the alien invasive fish in the river is predicted by the alternative model At the same time, the model prediction results are calculated based on the confusion matrix to calculate the following performance indicators: accuracy, F1 value, AUC value, etc.

[0060] ;

[0061] In the formula, TP is the number of true positives; TN is the number of true negatives; FP is the number of false positives; FN is the number of false negatives; TPR is the true positive rate; FPR is the false positive rate;

[0062] To comprehensively reflect the discrimination ability and classification stability of the model, a model comprehensive evaluation function is constructed :

[0063] ; In the formula, represents the macro-average area under the receiver operating characteristic curve; represents the macro-average F1 score; take The alternative model with the maximum value is selected as the invasion intensity prediction model. If the values of different models are similar, the model with fast reasoning speed and good interpretability is preferentially selected according to the model complexity, running efficiency and result stability. After determining the invasion intensity prediction model , the algorithm type, parameter combination, cross-validation configuration, random seed and weight file are solidified and output to ensure the consistency of subsequent interpretation and retraining.

[0064] (3) The invasion intensity prediction model is combined with the SHAP analysis algorithm to establish an invasion analysis model, the invasion response type and the invasion response priority of the native aquatic biota community to the alien invasive fish are analyzed by the invasion analysis model, the corresponding ecological regulation strategy is generated according to the invasion response type, and the implementation priority of the ecological regulation strategy is marked according to the invasion response priority.

[0065] The invasion analysis model is expressed as:

[0066] , ; In the formula: is the predicted result output by the invasion intensity prediction model , for feature marginal contribution to prediction result, global importance for measuring feature average contribution to overall prediction result, intra-class contribution for measuring feature prediction contribution to different class prediction results.

[0067] Specifically, global importance and intra-class contribution The results respectively represent the direction and intensity of each community feature in the prediction of the invasion gradient of alien fish, based on which, the analysis of the response type and priority of the native aquatic community to the invasion of alien fish is carried out:

[0068] (31) Response type determination

[0069] Calculate the weighted average contribution of each feature under all samples ; The response type of the feature is promotion response; The response type of the feature is inhibition response. Wherein, ; In the formula, is the weight of different class prediction results.

[0070] According to the above results, a response type table is established to automatically label the biological features of different communities with "promotion" or "inhibition" attributes. The determination process is automatically executed by an algorithm module without human intervention.

[0071] (32) Response priority determination

[0072] Calculate the community priority score , and determine the invasion response priority according to the priority score ; wherein, ; In the formula, is an adjustment coefficient, and its value range is 0.5-1.

[0073] Adjustment coefficient According to the direction stability index of the model explanation result, it is automatically calculated as follows:

[0074] ;

[0075] In the formula, is the feature intra-class contribution variance under different class prediction results, is the absolute value of the overall average contribution of the feature, The smoothing constant is used to prevent small positive numbers from being introduced due to numerical instability or division-by-zero errors when the denominator approaches zero. The typical value range is 10. -6 ~10 -4 The specific value is automatically set based on the order of magnitude of the feature contribution. When the overall model has high consistency, The value is close to 1; when the direction fluctuates greatly, The value is set close to 0.5 to dynamically adjust the weight of the response direction in the community priority calculation. This parameter design ensures that the priority score reflects both the importance of community indicators and the reliability and directional consistency of the model's interpretation results, thereby achieving the scientific rigor and robustness of the ranking of remediation and control strategies.

[0076] (33) Generation of ecological regulation strategies

[0077] The weakening regulatory strategies that promote response generation include, but are not limited to: controlling the population density of invasive fish species, reducing eutrophic load, and enhancing heterogeneity in habitat-single areas.

[0078] The enhanced regulatory strategies generated based on the described inhibition response include, but are not limited to: restoring emergent and submerged plant communities and enhancing plant diversity; optimizing substrate structure and microhabitat conditions to promote the recovery of benthic invertebrates; and enhancing algal community diversity and spatial cover to strengthen ecosystem stability.

[0079] Furthermore, to achieve systematic ecological management, this invention is further based on SHAP second-order interaction values. Establish community-based collaborative regulation pathways for any two characteristics. , Its SHAP second-order interaction value Defined as:

[0080]

[0081] Features in model interpretation When a feature is used in contribution calculation, the feature Marginal contribution to the prediction results;

[0082] The explanation does not include features. At that time, characteristics Editing contributions to the prediction results;

[0083] and Same as above, the above and The first-order SHAP value of the marginal contribution of the feature defined in step (3) of the application is consistent.

[0084] When , it indicates that the joint change of the features and promotes the invasion (joint promotion response), and correspondingly, a joint weakening regulation strategy should be generated and implemented; when , it indicates that the two communities have resistance or restraint, and correspondingly, a joint strengthening regulation strategy should be generated and implemented to maintain the synergistic relationship.

[0085] In summary, the final analysis result of the invasion regulation is output in the form of a "river invasion management and restoration suggestion table". The content includes: community name and key ecological indicators; response type (promotion / inhibition); regulation category (weakening / strengthening); suggested ecological measures and implementation priority. This directly provides a reference for river ecological restoration and alien fish control decision-making.

[0086] In addition, after implementing the control measures, the system can receive new monitoring data, automatically recalculate the SHAP contribution and community response direction, and realize dynamic updating. When the community structure or the abundance of alien fish changes, the model will automatically retrain and update the management suggestions, forming a self-adaptive closed-loop mechanism of "prediction-regulation-monitoring-re-prediction".

[0087] The application also discloses an invasion regulation system based on SHAP evaluation of alien fish invasion mechanism provided by an example embodiment, and the system comprises:

[0088] A data acquisition and processing layer is used for monitoring biological community type data and biological feature data of each biological community in a target river network region, and is also used for preprocessing the monitored data; specifically, six types of aquatic biological community indicators (fish, plankton, phytoplankton, sessile algae, aquatic plants, and benthic invertebrates) and related environmental characteristics thereof are included.

[0089] A model layer is based on the global importance and the in-class contribution degree output by an optimal classification model, and the results respectively represent the direction and intensity of each community feature in the prediction of the alien fish invasion gradient.

[0090] A decision layer is used for determining the invasion response type and the invasion response priority of the native aquatic biological community to the alien invasive fish according to the output result of the model layer, generating a corresponding ecological regulation strategy according to the invasion response type, and marking the implementation priority of the ecological regulation strategy according to the invasion response priority.

[0091] The invasion response type includes:

[0092] independent promotion response and independent inhibition response to a single feature;

[0093] joint promotion response and joint inhibition response to at least two features;

[0094] The ecological regulation strategy comprises:

[0095] independent weakening regulation strategy generated according to the independent promotion response and independent strengthening regulation strategy generated according to the independent inhibition response;

[0096] joint weakening regulation strategy generated according to the joint promotion response and joint strengthening regulation strategy generated according to the joint inhibition response.

[0097] As to the system in the above embodiments, the specific manner in which each structural layer performs the operation has been described in detail in the embodiments related to the method, and thus will not be described in detail here.

[0098] To sum up, the invasion regulation method and system for evaluating the invasion mechanism of alien fish species based on SHAP provided in the embodiment analyze and evaluate the influence mechanism of river alien invasive fish on native aquatic biological community through SHAP explainable algorithm, and map the threshold / contribution to control classification and management priority, realizing quantitative execution from algorithm explanation to ecological control; through identifying the response direction and interaction of different trophic level communities to alien invasive fish, quantitative basis is provided for river restoration and invasion prevention and control; an automatically updated community level management system is established, improving the scientificity and adaptability of river ecosystem management; subjective bias caused by traditional experience judgment and single factor decision is effectively avoided, realizing data-driven precise ecological regulation.

[0099] In another exemplary embodiment, an electronic device is also provided, which includes a memory and a processor, and a program stored in the memory, and the processor implements one or more steps of the foregoing method for evaluating the invasion mechanism of alien fish species based on SHAP when executing the program.

[0100] In another exemplary embodiment, a computer readable storage medium including program instructions is also provided, and the program instructions are executed by a processor to implement the steps of the foregoing method for evaluating the invasion mechanism of alien fish species based on SHAP. For example, the computer readable storage medium can be a first memory including program instructions, and the foregoing program instructions can be executed by a first processor of an electronic device to complete the foregoing method for evaluating the invasion mechanism of alien fish species based on SHAP.

[0101] In another exemplary embodiment, there is also provided a computer program product comprising a computer program capable of being executed by a programmable device, the computer program having code portions for performing the above-described method for SHAP-based evaluation of mechanisms of invasion of alien fish species for the purpose of invasion control when executed by the programmable device. In some embodiments, part or all of the computer program can be loaded and / or installed onto the device via a ROM and / or a communication unit. When the computer program is loaded into the RAM and executed by the CPU, one or more steps of the aforementioned method can be performed. Alternatively, in other embodiments, the CPU can be configured, by any other suitable means (e.g. by means of firmware), to perform one or more steps of the aforementioned method for SHAP-based evaluation of mechanisms of invasion of alien fish species for the purpose of invasion control.

[0102] It is to be noted that the above merely describes preferred embodiments of the present application and the principles of the technology applied. Those skilled in the art will understand that the present application is not limited to the specific embodiments described herein, and that various obvious changes, modifications and substitutions can be made thereto without departing from the scope of the present application. Therefore, although the present application has been described in detail by means of the above embodiments, the present application is not limited to the above embodiments but includes other equivalent embodiments, and the scope of the present application is defined by the appended claims.

Claims

1. An invasion control method based on SHAP for assessing the invasion mechanism of alien fish, characterized in that, Includes the following steps: (1) Data monitoring; monitoring data on the types of biological communities and the biological characteristics of each biological community within the target river network area; (2) Select an appropriate model to construct an intrusion intensity prediction model; Input the data monitored in step (1) into the candidate models, and calculate the accuracy, F1 score, AUC score, and comprehensive evaluation function value of each candidate model. Choose a suitable model; Evaluation function: In the formula, This represents the area under the macro-mean receiver operating characteristic curve; This represents the macro's average F1 score; (3) The invasion intensity prediction model is combined with the SHAP analysis algorithm to establish an invasion analysis model. The invasion analysis model is used to analyze the invasion response type and priority of the native aquatic community to the invasive fish species. The corresponding ecological regulation strategy is generated according to the invasion response type. The implementation priority of the ecological regulation strategy is marked according to the invasion response priority. in: The intrusion response types include facilitating responses and suppressing responses; The ecological regulation strategy includes a weakening regulation strategy generated based on the promoting response and a strengthening regulation strategy generated based on the inhibiting response.

2. The invasion control method based on SHAP for assessing the invasion mechanism of alien fish according to claim 1, characterized in that: The data on the types of biological communities include fish, zooplankton, phytoplankton, macrobenthic invertebrates, aquatic higher plants, and attached algae; The biometric data includes density. Biomass Number of species Shannon diversity Pielou uniformity Margalef richness .

3. The invasion control method based on SHAP for assessing the invasion mechanism of alien fish according to claim 1, characterized in that, Data preprocessing is also included before inputting the data monitored in step (1) into the candidate model: Undetected biometric data are set to zero; Logarithmic transformation and robust scaling are performed on the detected biometric data; In the formula, These are the original values ​​after logarithmic transformation; These are the standardized values ​​after robust scaling. This is the median of the entire population of variables; The interquartile range of the variable.

4. The invasion control method based on SHAP assessment of invasive alien fish mechanisms according to claim 1, characterized in that, In step (2), the accuracy, F1 score, and AUC score are calculated using the following formulas: ； In the formula, TP represents the number of true positives; TN represents the number of true negatives; FP represents the number of false positives; FN represents the number of false negatives; TPR represents the true positive rate; and FPR represents the false positive rate.

5. The invasion control method based on SHAP assessment of invasive alien fish mechanisms according to claim 1, characterized in that, The intrusion analysis model is expressed as follows: , In the formula: The prediction results output by the intrusion intensity prediction model , Features Marginal contribution to the prediction results The total number of features, global importance Used to measure characteristics Average contribution to the overall prediction results, intra-class contribution Used to measure characteristics The predictive contribution of different categories of prediction results.

6. The invasion control method based on SHAP assessment of invasive alien fish mechanisms according to claim 5, characterized in that, The steps for analyzing the invasion response types of native aquatic communities to invasive fish species using the invasion analysis model include: Calculate the weighted average contribution of the target feature across all samples. ; The intrusion response type of the target feature is an facilitating response; The intrusion response type of the target feature is a suppression response; in, In the formula, Weights for prediction results of different categories.

7. The invasion control method based on SHAP for assessing the invasion mechanism of alien fish according to claim 6, characterized in that, The intrusion response types include: Independent promoting and independent inhibiting responses for a single feature; Jointly promoting and jointly inhibiting responses for at least two characteristics.

8. The invasion control method based on SHAP assessment of invasive alien fish mechanisms according to claim 7, characterized in that, The ecological regulation strategies include: The independent weakening regulation strategy generated based on the independent promoting response and the independent strengthening regulation strategy generated based on the independent inhibiting response; The joint weakening control strategy generated based on the joint promoting response and the joint strengthening control strategy generated based on the joint inhibiting response.

9. The invasion control method based on SHAP assessment of invasive alien fish mechanisms according to claim 6, characterized in that, The steps for analyzing the invasion response priority of native aquatic communities to invasive fish species using the invasion analysis model include: Calculate community priority score And score according to the priority. Determine the intrusion response priority; among which, In the formula, The adjustment factor has a value range of 0.5 to 1.

10. The invasion control method based on SHAP for assessing the invasion mechanism of alien fish according to claim 9, characterized in that: The adjustment coefficient Calculated using the following formula: In the formula, Features Variance of within-class contribution under different category prediction results The absolute value of the contribution of the characteristic population average. This is the smoothing constant.

Images