Stable isotope mixing model performance evaluation method, device, equipment and medium

Abstract

The invention provides a performance evaluation method for a stable isotope mixing model, which comprises the following steps of: collecting and establishing a pollutant source information base containing various types of indexes, determining a standard mixing ratio of each pollution source based on the information base, and creating at least two groups of different prior information sets according to the standard mixing ratio, obtaining various types of indexes in the corresponding mixture according to a standard mixing ratio, further respectively constructing a plurality of data distribution models with various types of indexes, and generating data sets with different sample sizes for each data distribution model, combining the generated data set with corresponding prior information, inputting the combined data set into a to-be-evaluated stable isotope mixing model for operation, recording a test result, and comparing the test result with a standard mixing ratio to obtain the test accuracy of the stable isotope mixing model so as to evaluate the performance of the used model; the method is helpful for systematically evaluating the influence of different factors on a model result, and provides a solid scientific basis for accurate traceability of drainage basin pollution.

Description

Technical Field

[0001] The present application relates to the technical field of data processing, and particularly relates to a method, device, equipment and medium for evaluating the performance of a stable isotope mixing model. Background Art

[0002] In recent years, with the increasing attention in the field of environmental ecology to the problem of the sources of complex pollutants, the stable isotope analysis technology has developed rapidly. Based on the fractionation phenomenon of elements in biogeochemical processes, by measuring the stable isotope ratios in different environmental and biological samples, this technology can effectively infer their origins, historical processes and environmental conditions. Stable isotope analysis not only provides an important means for exploring ecosystem energy flow, climate change history, and biogeochemical cycles, etc., but also has a wide range of applications, covering multiple levels from basic research to actual environmental protection.

[0003] The core of the stable isotope analysis technology lies in tracking the migration path of substances in nature through the changes in the stable isotope ratios of specific elements. With the in-depth understanding of natural processes, researchers have gradually combined the stable isotope analysis technology with mathematical modeling to describe the distribution and variation laws of isotopes in different biogeochemical processes. On this basis, a series of general stable isotope mixing models have been formed, such as MixSIAR (Bayesian Mixing Models in R) and SIMMR (Stable Isotope Mixing Models in R), and these models are widely used in the source analysis of substances in different environmental media.

[0004] The value of the stable isotope mixing model lies in that it can provide direct evidence about the sources and processes of substances, which makes it an important tool for identifying and tracking the sources of pollutants. In addition, such models can also effectively evaluate the impacts of different pollution sources on the environmental system, providing key scientific basis for pollution control and environmental restoration. However, in the actual application process, the stable isotope mixing model faces multiple challenges, including problems such as source uncertainty, tracer type uncertainty, mixed data structure uncertainty, and model category uncertainty, etc., and these problems may significantly affect the result accuracy of the model. Summary of the Invention

[0005] In view of the above-mentioned disadvantages of the prior art, the present invention provides a method, device, equipment and medium for evaluating the performance of a stable isotope mixing model to solve the above technical problems.

[0006] The present invention provides a method for evaluating the performance of a stable isotope mixing model. The method includes: collecting and establishing a pollutant source information database, which contains various types of indicators for distinguishing different pollution sources; determining the standard mixing ratios of each pollution source based on the pollutant source information database, and creating at least two different prior information sets based on the standard mixing ratios; obtaining the various types of indicators in the corresponding mixture according to the standard mixing ratios of each pollution source, constructing multiple data distribution models with various types of indicators respectively according to the various types of indicators of the mixture as expectations or intermediate values, and generating multiple data sets with different sample sizes for each data distribution model; inputting the generated data sets combined with the corresponding prior information into the stable isotope mixing model to be evaluated for running, and recording the test results of each stable isotope mixing model to be evaluated; comparing the test results with the standard mixing ratios to obtain the test accuracy of the stable isotope mixing model, so as to evaluate the performance of the used stable isotope mixing model.

[0007] In an embodiment of the present invention, after inputting the generated data sets combined with the corresponding prior information into the stable isotope mixing model to be evaluated for running, it further includes: defining a combination of any data set and prior information as a test scenario; respectively recording the running times of each stable isotope mixing model in different scenarios, and calculating the average running time of multiple stable isotope mixing models in the same scenario; comparing the running time of any stable isotope mixing model with the average running time to obtain the working efficiency of each stable isotope mixing model in different scenarios.

[0008] In an embodiment of the present invention, evaluating the performance of the used stable isotope mixing model includes: respectively testing the test accuracy and working efficiency of different stable isotope mixing models in the same test scenario; converting the test accuracy and the working efficiency into accuracy scores and efficiency scores respectively based on a preset conversion rule, and performing weighted calculation on the accuracy scores and the efficiency scores based on a preset weight ratio to obtain the performance scores of each stable isotope mixing model in the test scenario; traversing all test scenarios to obtain the performance scores of each stable isotope mixing model in different test scenarios, so as to evaluate the performance of each model.

[0009] In an embodiment of the present invention, comparing the test results with the standard mixing ratio to obtain the test accuracy of the stable isotope mixing model includes: determining the standard contribution ratio of each pollution source based on the standard mixing ratio, obtaining various types of indicators in the corresponding mixture according to the standard contribution ratio of the pollution source, constructing multiple data sets with different sample size data distributions having multiple types of indicators respectively with the various types of indicators of the mixture as expectations or intermediate values, and obtaining the test contribution ratio of each pollution source based on the test results of the stable isotope mixing model; calculating the difference between the standard contribution ratio and the test contribution ratio corresponding to the same pollution source to obtain the test error of the stable isotope mixing model used for each pollution source; and comprehensively calculating the test errors of different pollution sources to obtain the test accuracy of the stable isotope mixing model used.

[0010] In an embodiment of the present invention, comprehensively calculating the test errors of different pollution sources includes at least one of the following: calculating the corresponding root mean square error based on the test errors of each pollution source; calculating the corresponding mean absolute percentage error based on the test errors of each pollution source.

[0011] In an embodiment of the present invention, the method further includes: setting the number of pollution sources, the types of tracers, and the number of tracers; randomly combining the number of pollution sources, the types of tracers, and the number of tracers, and determining each combination result as a test condition to generate multiple test scenarios under different test conditions.

[0012] In an embodiment of the present invention, the prior information set at least includes a standard information set and an interference information set, wherein the data mixing ratio of the standard information set satisfies the standard mixing ratio, and the data mixing ratio of the interference information set deviates from the standard mixing ratio; the data distribution model at least includes a normal distribution, a uniform distribution, and a discrete distribution.

[0013] The present application provides a device for evaluating the performance of a stable isotope mixing model. The device includes: an information collection module for collecting and establishing a pollutant source information database, which contains various types of indicators for distinguishing different pollution sources; a test preparation module for determining the standard mixing ratios of each pollution source based on the pollutant source information database and creating at least two different prior information sets based on the standard mixing ratios; a test data generation module for respectively constructing multiple data distribution models according to the prior information sets and generating multiple data sets with different sample sizes for each data distribution model; a test implementation module for inputting the generated data sets combined with the corresponding prior information into the stable isotope mixing model to be evaluated for running and recording the test results of each stable isotope mixing model to be evaluated; and a model performance evaluation module for comparing the test results with the standard mixing ratios to obtain the test accuracy of the stable isotope mixing model so as to evaluate the performance of the used stable isotope mixing model.

[0014] The present application provides an electronic device. The electronic device includes: one or more processors; a storage device for storing one or more programs, which, when executed by the one or more processors, cause the electronic device to implement the method for evaluating the performance of a stable isotope mixing model as described above.

[0015] The present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor of a computer, the computer is caused to execute the method for evaluating the performance of a stable isotope mixing model as described above.

[0016] Advantages of the present invention: The method for evaluating the performance of a stable isotope mixing model proposed in the present invention collects and establishes a detailed pollutant source information database, which includes various types of indicators for distinguishing different pollution sources. When selecting a model, a more comprehensive and accurate data basis can be used, ensuring that the selected model can better reflect the actual situation. At least two different prior information sets are created based on the standard mixing ratios, and the various types of indicators in the corresponding mixtures are obtained according to the standard mixing ratios of each pollution source. Multiple data distribution models with various types of indicators are constructed respectively based on the various types of indicators of the mixtures as expectations or intermediate values. Datasets with different sample sizes are generated for each data distribution model, which can not only cover a wide range of actual application scenarios but also effectively reduce the result deviation caused by a single data distribution or sample size, thereby enhancing the reliability of the final result. Through the simulation analysis of different data distribution forms and sample size sizes, it can be clarified under which conditions a specific model performs best, which helps the effective allocation of resources. By comparing the test results with the standard mixing ratios, the test accuracy of the stable isotope mixing model is obtained, providing solid data support for decision-making in actual applications. Generally speaking, this method provides a standardized evaluation system that can be used to compare the performance differences between different stable isotope mixing models. In addition, by systematically evaluating the performance of the stable isotope mixing model under different conditions, not only the scientificity and accuracy of model selection are improved, but also the reliability and practicality of the results are enhanced, providing strong support for solving complex environmental pollution problems.

[0017] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and do not limit this application. Brief Description of the Drawings

[0018] The drawings herein are incorporated into the specification and constitute a part of this specification, showing embodiments consistent with this application, and are used together with the specification to explain the principles of this application. Obviously, the drawings in the following description are only some embodiments of this application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts. In the drawings:

[0019] Figure 1 is a schematic diagram of the implementation environment of the method for evaluating the performance of a stable isotope mixing model shown in an exemplary embodiment of this application;

[0020] Figure 2 is a flowchart of the implementation of the method for evaluating the performance of a stable isotope mixing model shown in an exemplary embodiment of this application;

[0021] Figure 3a is the influence of the number of sources and the presence or absence of prior proportion information on the model accuracy shown in an exemplary embodiment of this application;

[0022] Figure 3b The influence of different tracer types and numbers on the model performance shown in an exemplary embodiment of the present application;

[0023] Figure 4a The influence of different numbers of mixture samples on the model performance shown in an exemplary embodiment of the present application;

[0024] Figure 4b The change in model performance under different data distribution structures shown in an exemplary embodiment of the present application;

[0025] Figure 5a The running time of the MixSIAR model under different scenarios shown in an exemplary embodiment of the present application;

[0026] Figure 5b The running time of the SIMMR model under different scenarios shown in an exemplary embodiment of the present application;

[0027] Figure 6 The block diagram of a stable isotope mixing model performance evaluation device shown in an exemplary embodiment of the present application;

[0028] Figure 7 The structural schematic diagram of a computer system of an electronic device suitable for implementing the embodiments of the present application is shown. Detailed implementation manners

[0029] The following will illustrate the implementation manners of the present invention with reference to the drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific implementation manners. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention, rather than for limiting the protection scope of the present invention.

[0030] It should be noted that the drawings provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner. Therefore, only the components related to the present invention are shown in the drawings, rather than being drawn according to the number, shape, and size of the components in actual implementation. The type, number, and ratio of each component in actual implementation can be arbitrarily changed, and the component layout type may also be more complex.

[0031] In the following description, numerous details are explored to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present invention difficult to understand.

[0032] First of all, it should be noted that Stable Isotope Mixing Models are a mathematical tool based on stable isotope analysis technology, used to estimate the relative contribution ratios of different sources to a certain mixture (such as pollutants, nutrients, etc.) in an ecosystem. These models infer their origins, historical processes, and environmental conditions by measuring and analyzing the stable isotope ratios of specific elements in different environmental media.

[0033] MixSIAR (Bayesian Mixing Models in R) is a stable isotope mixing model based on the Bayesian statistical framework, mainly used to analyze the contribution ratios of different sources to a certain mixture in an ecosystem.

[0034] SIMMR (Stable Isotope Mixing Models in R) is another R package for stable isotope mixing analysis, aiming to simplify and accelerate the application of stable isotope mixing models.

[0035] Figure 1 It is a schematic diagram of the implementation environment of the stable isotope mixing model performance evaluation method shown in an exemplary embodiment of this application.

[0036] As Figure 1 shown, the implementation environment of the stable isotope mixing model performance evaluation method includes a data acquisition module 101 and a computer device 102. Among them, the data acquisition module is used to collect necessary tracer indicators from different pollution sources, and the indicators include but are not limited to stable isotopes (such as δ 13 C-DOC, δ 15 N-NO3 - -N), ultraviolet-visible spectrum indicators, fluorescence spectrum indicators, and stoichiometric ratios, etc. In addition, it also involves the analysis of various environmental source samples such as fallen leaves, riparian plants, algae, soil, and groundwater. Specifically, the data set generated by the data acquisition module contains detailed pollutant source information, which is used to construct a pollutant source information library and determine the standard mixing ratios of each pollution source. In addition, the data acquisition module is also used to create at least two different sets of prior information based on the standard mixing ratios.

[0037] The computer device 102 is used to receive data from the data acquisition module 101 and perform further processing and analysis on this basis. Specifically, the computer device 102 will construct multiple data distribution models (such as normal distribution, uniform distribution, and discrete distribution) based on the received data, and generate data sets with different sample sizes for each model. Subsequently, the generated data sets are combined with the corresponding prior information and then input into the stable isotope mixing model to be evaluated (such as MixSIAR or SIMMR) for operation.

[0038] It can be understood that the implementation environment of the stable isotope mixing model performance evaluation method realizes the full-process automated management from data collection, processing to model operation and performance evaluation through the effective cooperation of the data acquisition module 101 and the computer device 102, providing strong technical support for accurately identifying and tracing the sources of pollutants.

[0039] Figure 2 It is the implementation flowchart of the stable isotope mixing model performance evaluation method shown in an exemplary embodiment of the present application.

[0040] As Figure 2 shown, in an exemplary embodiment, the stable isotope mixing model performance evaluation method at least includes steps S210 to S250, which are introduced in detail as follows:

[0041] Step S210, collect and establish a pollutant source information database, and the pollutant source information database contains various types of indicators for distinguishing different pollution sources.

[0042] In order to distinguish different pollution sources and effectively trace the sources of pollutants, the common pollution sources in the basin and their corresponding tracer indicators are first determined.

[0043] In an embodiment of the present application, the pollution sources include but are not limited to fallen leaves, riparian plants, algae, soil, groundwater, pig manure sewage, dairy livestock wastewater, agricultural wastewater, industrial facility wastewater, urban sewage, aquaculture treatment facility wastewater, and wastewater treatment facility effluent. For each pollution source, detailed tracer data are collected respectively, including dissolved organic carbon (DOC), dissolved organic nitrogen (DON), aromaticity index (SUVA 254 ), decay index (HIX), fluorescence index (FI), autochthonous index (BIX), stable isotope ratio (such as δ 13 C-DOC, δ 15 N-NO3 - -N), and stoichiometric ratio (C / N). It should be noted that during the sample collection process, it is necessary to ensure the consistency and repeatability of data collection for each selected pollution source, record the specific location, time, weather conditions, etc. of the sampling site, and analyze the samples using standard methods.

[0044] Subsequently, a detailed pollutant source information database is constructed based on the collected data. This information database not only includes the specific tracer measurement values of various pollution sources, but also integrates the characteristic descriptions of each pollution source, the map identification of the collection locations, and the relevant environmental background information, so as to obtain a database that can comprehensively reflect the characteristics of different pollution sources within the basin, providing a solid basic dataset for subsequent model input.

[0045] Step S220: Determine the standard mixing ratios of each pollution source based on the pollutant source information database, and create at least two different prior information sets based on the standard mixing ratios.

[0046] In an embodiment of the present application, before determining the standard mixing ratios of each pollution source based on the pollutant source information database, the method further includes: setting the number of pollution sources, the types of tracers, and the number of tracers; randomly combining the number of pollution sources, the types of tracers, and the number of tracers, and determining each combination result as a test premise to generate multiple test scenarios under different test premises.

[0047] To determine the standard mixing ratios of each pollution source based on the pollutant source information database, first set the number of pollution sources, the types of tracers, and their quantities. For example, in a specific implementation, 8 main pollution sources (such as plants, soil, groundwater, pig and livestock sewage, agricultural wastewater, industrial wastewater, urban sewage, and effluent from sewage treatment plants) are selected, and 7 tracers (DOC / DTN, SUVA 254 , HIX, FI, BIX, δ 13 C-DOC, δ 15 N-NO3 - ) are selected for each pollution source. Then, these settings are randomly combined through a programming tool (such as the R language), a certain number are randomly selected from the selected pollution sources and tracers for combination each time, and each such combination is regarded as a test premise, so as to generate multiple test scenarios to comprehensively evaluate the model performance under different conditions.

[0048] Step S230: Obtain various types of indicators in the corresponding mixture according to the standard mixing ratios of each pollution source, respectively construct multiple data distribution models with various types of indicators for the mixture, and generate multiple datasets with different sample sizes for each data distribution model.

[0049] In an embodiment of the present application, the prior information set at least includes a standard information set and an interference information set, wherein the data mixing ratio of the standard information set satisfies the standard mixing ratio, and the data mixing ratio of the interference information set deviates from the standard mixing ratio; the data distribution models at least include a normal distribution, a uniform distribution, and a discrete distribution.

[0050] It should be specifically noted that in the actual operation process, it is very difficult to adjust the data to be exactly the same as the preset standard mixing ratio. Therefore, the "meeting" mentioned in this embodiment does not mean that the data mixing ratio of the standard information set must be exactly the same as the standard mixing ratio. On the contrary, it means that the data mixing ratio should basically conform to the standard mixing ratio, and the deviation degree should be within an acceptable range. This flexibility ensures that in practical applications, even if there are certain deviations, the effectiveness and reliability of model evaluation can be guaranteed.

[0051] In a specific embodiment of the present application, after determining the standard mixing ratio of each pollution source, a prior information set is created for each test premise. This set includes at least two parts: the standard information set and the interference information set. The data mixing ratio in the standard information set is generally set within the range of ±5% of the preset standard mixing ratio. For example, if the standard mixing ratio is 30% for fallen leaves, 20% for soil, 10% for groundwater, 25% for livestock sewage, and 15% for urban sewage, then the data mixing ratio in the standard information set is within the range of this ratio ±5%. The data mixing ratio in the interference information set is intentionally deviated from the standard mixing ratio to create non-ideal prior information for testing the sensitivity of the model to outliers or uncertainties.

[0052] In addition, after setting the combination of pollution sources and tracers and the prior information, the standard mixing ratio of each pollution source is applied to obtain various types of indicators in the corresponding mixture. The numerical values of these indicators are used as expectations or intermediate values to form datasets with multiple types of indicators and different sample size data distribution models. For example, three forms of normal distribution, uniform distribution, and discrete distribution are adopted to generate datasets with 50 and 100 sample points respectively. In this way, the influence of different data structures on the model performance can be explored. In the R program, scripts are written to automatically execute the above steps: automatically generate a specified number of combinations of pollution sources and tracers, determine the standard mixing ratio of each pollution source, generate the corresponding prior information set accordingly, and generate the corresponding simulated datasets according to different data distribution models generated by using various types of indicators in the corresponding mixture generated by the standard mixing ratio of each pollution source as expectations or intermediate values. These datasets are then input into the MixSIAR or SIMMR model for analysis. By comparing the differences (such as RMSE, MAPE) between the model output results and the actual mixing ratio under different test scenarios, the performance of the model under different conditions can be evaluated.

[0053] Step S240: After combining the generated dataset with the corresponding prior information, input it into the stable isotope mixing model to be evaluated and run, and record the test results of the contribution ratio of each pollution source.

[0054] In one embodiment of the present application, in order to verify the performance of stable isotope mixing models (such as MixSIAR and SIMMR) under different conditions, after creating a prior information set containing a standard information set and an interference information set for each test scenario, the generated data sets are combined with the corresponding prior information and then input into a pre-set stable isotope mixing model for running. This process is automatically executed using an R language script. Open and run a file named main.R in the RStudio environment. This file calls other necessary R files, including analysis.R, dataprocess.R, etc., to ensure seamless connection of data processing and model running. For each data set and prior information combination, run the MixSIAR and SIMMR models with and without prior information respectively, and record the contribution ratio of each pollution source and the model running time. For example, in a specific test case, when testing with 50 discrete distribution mixture data generated by combining 5 pollution sources and all tracer combinations 1, the output results of the MixSIAR and SIMMR models are recorded with and without the prior standard information set.

[0055] Step S250, compare the test results with the standard mixing ratio to obtain the test accuracy of the stable isotope mixing model, so as to evaluate the performance of the used stable isotope mixing model.

[0056] In one embodiment of the present application, comparing the test results with the standard mixing ratio to obtain the test accuracy of the stable isotope mixing model includes: determining the standard contribution ratio of each pollution source based on the standard mixing ratio, and obtaining the test contribution ratio of each pollution source based on the test results of the stable isotope mixing model; calculating the difference between the standard contribution ratio and the test contribution ratio corresponding to the same pollution source to obtain the test error of the used stable isotope mixing model for each pollution source; comprehensively calculating the test errors of different pollution sources to obtain the test accuracy of the used stable isotope mixing model. Among them, comprehensively calculating the test errors of different pollution sources includes at least one of the following: calculating the corresponding root mean square error based on the test errors of each pollution source; calculating the corresponding mean absolute percentage error based on the test errors of each pollution source.

[0057] To evaluate the test accuracy of stable isotope mixing models (such as MixSIAR and SIMMR), first determine the standard contribution ratio of each pollution source based on a pre-set standard mixing ratio.

[0058] In a specific case, the standard mixing ratio is 30% for fallen leaves, 20% for soil, 10% for groundwater, 25% for livestock sewage, and 15% for urban sewage. Then, a stable isotope mixing model is used to analyze the generated dataset to obtain the test contribution ratios of each pollution source. The model is run through an R language script, and the output results are recorded. Continuing with the above example, assume the results predicted by the model are 28% for fallen leaves, 22% for soil, 9% for groundwater, 26% for livestock sewage, and 15% for urban sewage. Subsequently, the differences between the standard contribution ratios and the test contribution ratios of individual pollution sources are calculated respectively, and these are used as the test errors for that pollution source. For example, for fallen leaves, its test error is |30% - 28%| = 2%; similarly, the test errors for other pollution sources can be calculated. Next, in order to comprehensively evaluate the test accuracy of the entire model, the test errors of different pollution sources need to be calculated comprehensively. For this purpose, the following two main methods are adopted: one is to calculate the corresponding root mean square error (RMSE) based on the test errors of each pollution source, and the other is to calculate the mean absolute percentage error (MAPE). Specifically as follows:

[0059]

[0060] where n is the number of sources; O i and P i represent the actual and predicted contribution ratios respectively.

[0061] Finally, ultimately, based on the calculated RMSE and / or MAPE values, the overall test accuracy of the used stable isotope mixing model is evaluated. Smaller RMSE and MAPE values indicate higher accuracy and reliability of the model. For example, in a test scenario, if the calculated RMSE is 0.02 and the MAPE is 0.5, it means that the model performs ideally in this scenario and can better predict the contribution ratios of each pollution source.

[0062] In addition, in an embodiment of the present application, after the generated dataset is combined with the corresponding prior information and input into the stable isotope mixing model to be evaluated and run, it further includes: defining a combination of any dataset and prior information as a test scenario; respectively recording the running times of each stable isotope mixing model under different scenarios, and calculating the average running time of multiple stable isotope mixing models under the same scenario; comparing the running time of any stable isotope mixing model with the average running time to obtain the working efficiency of each stable isotope mixing model under different scenarios.

[0063] In a specific embodiment of the present application, in order to further evaluate the working efficiency of different stable isotope mixing models (such as MixSIAR and SIMMR), after combining the generated data set with the corresponding prior information and inputting it into a preset model for running, the running time of the model in each test scenario is recorded and analyzed.

[0064] First, a combination of any data set and prior information is defined as an independent test scenario. For example, if different combinations of 5 pollution sources and 7 tracers are generated, and three data distribution forms, namely normal distribution, uniform distribution, and discrete distribution, are considered, and data sets with 50 and 100 sample points are generated under each distribution, then each such combination is an independent test scenario.

[0065] In each test scenario, the corresponding combined data set and prior information are input into the stable isotope mixing model to be evaluated for running. When performing this process in the R language environment, in addition to recording the contribution ratio of each pollution source, the running times of the MixSIAR and SIMMR models with and without prior information are particularly recorded.

[0066] After completing the model running for all test scenarios, the obtained running times are summarized and analyzed. For each type of stable isotope mixing model (such as MixSIAR and SIMMR), its average running time in the same test scenario is calculated. Suppose that in a specific scenario, the running time of the MixSIAR model is 3 minutes and the running time of SIMMR is 5 minutes. By calculating, the average running duration of the two models in this scenario is 4 minutes, which indicates that in this scenario, the running efficiency of the MixSIAR model is higher than the average level, and the running efficiency of the SIMMR model is lower than the average level. Similarly, the testing methods for more models are the same as those of the above two models. Then, by traversing different test scenarios, the performance of each model in different test scenarios can be obtained, and further, the performance of different models in different test scenarios can be evaluated.

[0067] In an embodiment of the application, evaluating the performance of the stable isotope mixing model used includes: respectively testing the test accuracy and working efficiency of different stable isotope mixing models in the same test scenario; converting the test accuracy and working efficiency into accuracy scores and efficiency scores respectively based on preset conversion rules, and performing weighted calculations on the accuracy scores and efficiency scores based on a preset weight ratio to obtain the performance scores of each stable isotope mixing model in the test scenario; traversing all test scenarios to obtain the performance scores of each stable isotope mixing model in different test scenarios, so as to evaluate the performance of each model.

[0068] In a specific embodiment of the present application, based on a preset conversion rule, the test accuracy and work efficiency are converted into corresponding scores. For the accuracy score, an ideal value can be set (for example, both RMSE and MAPE are 0), and then the score is determined according to the gap between the actual test result and the ideal value. Suppose the ideal upper limit of RMSE is 0.05, then a model with an RMSE of 0.02 can obtain a relatively high accuracy score, such as 90 points (out of 100). For the efficiency score, it can also be evaluated based on the running time of the model, and a shorter running time gets a higher score. For example, a model with a running time of 35 seconds can obtain an efficiency score close to full marks, while a model with a running time of 4 minutes gets a lower score, such as 60 points (out of 100). It should be noted that the specific conversion rule is set and adjusted according to the requirements of the real-time application scenario, and the present application does not impose any specific restrictions on it.

[0069] Then, based on the preset weight ratio, the accuracy score and the efficiency score are weighted and calculated to obtain the performance scores of each stable isotope mixing model in the test scenario. Suppose the weight of the accuracy score is 70% and the efficiency score is 30%. In the above example, if the accuracy score of the MixSIAR model is 90 points and the efficiency score is 60 points, then its performance score is 81 points. For the SIMMR model, if its accuracy score is 92 points and the efficiency score is 95 points, then its performance score is 93.1 points. After that, all test scenarios are traversed to obtain the performance scores of each stable isotope mixing model in different test scenarios. For example, through the analysis of multiple test scenarios, it is found that the MixSIAR model performs better than the SIMMR model under certain complex data distributions, but in most cases, the SIMMR model obtains a higher comprehensive performance score due to its faster running time and relatively high accuracy score.

[0070] Finally, the performance scores in all test scenarios are summarized to evaluate the overall performance of each stable isotope mixing model. For example, the summary results show that the SIMMR model performs well in most test scenarios. Especially when quick response or handling a large amount of data is required, its high efficiency makes it a preferred solution. However, when facing an extremely complex pollution source identification task, the MixSIAR model is given priority consideration due to its higher accuracy.

[0071] In addition, it should be emphasized that in the above embodiments, usually, the combination of the number of pollution sources, the types of tracers and their quantities is first set as the test conditions, and then based on these conditions, an appropriate data mixing ratio, data model type (such as normal distribution, uniform distribution or discrete distribution), and data volume are selected to construct a specific test scenario. However, in practical applications, this process can be more flexible. Specifically, the data mixing ratio, data model type and data volume can be determined first, and then the number of pollution sources, the types of tracers and their quantities can be adjusted. This method is equally effective and does not necessarily strictly follow a certain fixed order for combination and adjustment. Regardless of which method is adopted, the ultimate goal is to generate diverse test scenarios based on different numbers of pollution sources, types of tracers, quantities of tracers, data mixing ratios, data model types and data volumes, and to evaluate the performance of each stable isotope mixing model (such as MixSIAR and SIMMR) under different scenarios respectively.

[0072] Therefore, in an embodiment of the present application, a method for evaluating the performance of a stable isotope mixing model includes the following steps:

[0073] Step 1, determine a set of test conditions, where the test conditions include the number of pollution sources, the types of tracers and their quantities;

[0074] Step 2, select a data mixing ratio, a data model type and a data volume according to the test conditions to generate multiple test scenarios;

[0075] Step 3, in each test scenario, input the corresponding data set and prior information into at least one stable isotope mixing model to be evaluated and run, and record the contribution ratio of each pollution source and the model running time;

[0076] Step 4, calculate the test accuracy and working efficiency of each stable isotope mixing model in each test scenario;

[0077] Step 5, based on a preset conversion rule, convert the test accuracy and working efficiency into an accuracy score and an efficiency score respectively;

[0078] Step 6, perform a weighted calculation on the accuracy score and the efficiency score according to a preset weight ratio to obtain the performance score of each stable isotope mixing model in each test scenario;

[0079] Step 7, traverse all test scenarios to obtain the performance scores of each stable isotope mixing model in different test scenarios, so as to evaluate the overall performance of each model;

[0080] Among them, the method allows flexible adjustment of the operation order, that is, the number of pollution sources, the types of tracers and their quantities can be adjusted after determining the data mixing ratio, the data model type and the data volume, or the combination and adjustment order of its data categories is not emphasized.

[0081] In a specific embodiment of the present application, taking the evaluation of the stable isotope mixing models MixSIAR and SIMMR as an example, the specific test scenario information is shown in Table 1 as follows:

[0082] Table 1

[0083]

[0084]

[0085] Input the test scenario information shown in Table 1 into the stable isotope mixing models MixSIAR and SIMMR respectively, and the corresponding test results are shown in Table 2 as follows:

[0086] Table 2

[0087]

[0088]

[0089]

[0090] Under the test scenario with scenario number 1, the test results of MixSIAR and SIMMR are shown in Figure 3. When the number of sources is reduced, both the RMSE and MAPE of the two models increase ( Figure 3a ), indicating that reducing the number of sources will reduce the accuracy of model operation. Regarding the prior information, there is no significant difference in the RMSE and MAPE between the MixSIAR model and the SIMMR model, indicating that the prior information cannot improve the accuracy of model operation ( Figure 3a ); for different tracer types and numbers, when using the spectral feature index as a tracer, compared with using only isotopes as tracers, both the RMSE and MAPE of the MixSIAR model and the SIMMR model show a significant increase. This indicates that it is not appropriate to use the spectral feature index alone as a tracer for organic matter source tracing. However, when using the spectral feature index, stable isotopes, and stoichiometric ratios in combination as tracers, both the RMSE and MAPE of the model show a decrease, indicating that the combined use of multiple tracers can improve the prediction accuracy of the model. At the same time, both the RMSE and MAPE of the MixSIAR model are significantly higher than those of the SIMMR model, indicating that when using other indicators as tracers in addition to stable isotopes, the output results of the SIMMR model are better than those of the MixSIAR model ( Figure 3b ). For different numbers of mixture data, the RMSE and MAPE of the two models are not much different, indicating that the change in the number of mixture data does not affect the model results ( Figure 4a ).

[0091] In the test scenario with a scenario number of 2 - 3 and a mixture data structure (50 data + 8 sources + 2 isotope tracers + no prior + normal distribution), the MixSIAR model and the SIMMR model did not show significant differences under normal distribution and discrete distribution data structures. However, when the data follows a uniform distribution, both RMSE and MAPE increase, indicating that when the data distribution is close to normal distribution, the performance of the model is often more stable and reliable. Figure 4b ) This shows that although the natural properties of stable isotopes themselves conform to the Gaussian distribution theorem, under different data structures, the performance of stable isotope models will show relatively significant fluctuations. Therefore, according to practical experience, a relatively large number of samples are required to meet the assumptions of the data distribution structure. At the same time, before the actual application of the model, it is recommended to use statistical methods such as data exploration and preprocessing techniques to verify the rationality of the data distribution assumptions. According to these analysis results, if necessary, the data can be appropriately transformed to enhance the applicability and prediction accuracy of the model.

[0092] In addition, through comprehensive analysis of the running time of the models in the aforementioned test scenarios, the test results shown in Figure 5 are obtained. In any case, the running time of the SIMMR model is shorter than that of the MixSIAR model. The former is measured in seconds, while the latter is counted in hours. When the number of mixture data is increased, the time required for both the SIMMR model and the MixSIAR model shows a multiple increase, but the MixSIAR model increases more, reaching about 10 times; and under different data distributions, the running time of the MixSIAR model varies greatly. Especially when the data adopts a uniform distribution, the running time can reach about 100h. Figure 5a ) However, the running time of the SIMMR model changes little. Figure 5b ) For different numbers of sources, the running time of both models changes little; for whether there is prior information during model operation, the running time of the SIMMR model changes little, but in the case of no prior information, the running time of the MixSIAR model can increase by up to 10 times; and in the case of different tracers, the more the number of tracers, the longer the running time of the model, and spectral indicators can significantly increase the running time of both models, especially the MixSIAR model.

[0093] Based on the above test results, it can be seen that the MixSIAR model can provide more detailed customized solutions under certain specific conditions (such as complex multi-pollution source environments), and is suitable for situations where multiple potential influencing factors need to be considered. However, its running time is relatively long, especially when dealing with uniformly distributed data or a large amount of data, and its efficiency is relatively low. The SIMMR model, on the other hand, can maintain high accuracy and extremely high computational efficiency in datasets with normal distribution, uniform distribution, or discrete distribution, and is particularly suitable for scenarios that require quick response and handling of large-scale datasets. However, it is not as flexible as MixSIAR when dealing with extremely complex pollution source identification tasks. Therefore, in practical applications, the appropriate model should be selected according to the specific environmental characteristics and requirements. For scenarios with scarce samples or the need for quick response, SIMMR is a better choice; while for complex and multi-level ecosystem analysis, MixSIAR may be more applicable.

[0094] Thus, based on the performance evaluation method of the stable isotope mixing model proposed in this application, a systematic method is provided to evaluate the performance of different stable isotope mixing models under various conditions. Specifically, this method constructs a comprehensive evaluation framework for the environmental pollution source identification model. By simulating different data distribution forms (normal distribution, uniform distribution, discrete distribution), adjusting the sample size (50 and 100 mixture data points), changing the number of sources, and adjusting the number and types of tracers, the impacts of these factors on the model performance are systematically evaluated. Its systematic evaluation framework ensures a comprehensive understanding of the model performance and helps to select the most suitable model for specific needs in practical applications. Moreover, this method not only quantifies the gap between the model prediction value and the actual value through the root mean square error (RMSE) and mean absolute percentage error (MAPE) to accurately measure the prediction ability of the model, but also records and compares the running times of each model. Especially when dealing with a large amount of data or complex scenarios, the computational efficiency becomes one of the key indicators, making the model performance evaluation dimension more abundant and effectively improving the credibility of its evaluation results.

[0095] In addition, it should be specifically noted that the embodiments of this application mainly take two stable isotope mixing models, MixSIAR and SIMMR, as examples for specific demonstration. However, in practical applications, this performance evaluation method is equally applicable to any other type of stable isotope mixing model. The specific evaluation steps are the same as those of the aforementioned models, and only corresponding adjustments need to be made according to the characteristics of the specific model. In view of this, the evaluation processes of other models will not be elaborated one by one in this article. It can be understood that the performance evaluation method for stable isotope mixing models proposed in this application is not limited to specific models, but provides a general framework that can be widely applied to the performance evaluation of various stable isotope mixing models, thus providing great convenience for researchers and practitioners when selecting and optimizing models. In this way, users can easily expand the application scope of this method according to their own needs and specific situations to ensure that the selected model can achieve the best effect in practical applications.

[0096] Figure 6 is a block diagram of a device for evaluating the performance of a stable isotope mixing model shown in an exemplary embodiment of this application. This device can be applied to Figure 1 the implementation environment shown. This device can also be applicable to other exemplary implementation environments and be specifically configured in other devices. This embodiment does not limit the implementation environment applicable to this device.

[0097] As Figure 6 shown, this exemplary device for evaluating the performance of a stable isotope mixing model includes: an information collection module 810, a test preparation module 820, a test data generation module 830, a test implementation module 840, and a model performance evaluation module 850.

[0098] Among them, the information collection module 810 is used to collect and establish a pollutant source information database, and the pollutant source information database contains various types of indicators for distinguishing different pollution sources; the test preparation module 820 is used to determine the standard mixing ratios of each pollution source based on the pollutant source information database and create at least two different prior information sets based on the standard mixing ratios; the test data generation module 830 is used to obtain the various types of indicators in the corresponding mixture according to the standard mixing ratios of each pollution source, construct multiple data distribution models with various types of indicators respectively according to the various types of indicators of the mixture, and generate multiple data sets with different sample sizes for each data distribution model; the test implementation module 840 is used to input the generated data sets combined with the corresponding prior information into the stable isotope mixing model to be evaluated for running and record the test results of each stable isotope mixing model to be evaluated; the model performance evaluation module 850 is used to compare the test results with the standard mixing ratios to obtain the test accuracy of the stable isotope mixing model so as to evaluate the performance of the used stable isotope mixing model.

[0099] It should be noted that the stable isotope mixing model performance evaluation device provided in the above embodiments and the stable isotope mixing model performance evaluation method provided in the above embodiments belong to the same concept. The specific ways in which each module and unit perform operations have been described in detail in the method embodiments and will not be elaborated here. In practical applications, the stable isotope mixing model performance evaluation device provided in the above embodiments can, according to needs, allocate the above functions to different functional modules, that is, divide the internal structure of the device into different functional modules to complete all or part of the functions described above. This is not limited here either.

[0100] An embodiment of the present application further provides an electronic device, including: one or more processors; a storage device for storing one or more programs, and when the one or more programs are executed by the one or more processors, the electronic device realizes the stable isotope mixing model performance evaluation method provided in each of the above embodiments.

[0101] Figure 7 The structural schematic diagram of a computer system of an electronic device suitable for implementing the embodiments of the present application is shown. It should be noted that Figure 7 The computer system 700 of the electronic device shown is only an example and should not impose any limitation on the functions and usage scope of the embodiments of the present application.

[0102] As Figure 7 shown, the computer system 700 includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 702 or the program loaded from the storage section 708 into the random access memory (RAM) 703, such as executing the method described in the above embodiments. In the RAM 703, various programs and data required for system operation are also stored. The CPU 701, ROM 702, and RAM 703 are connected to each other through a bus 704. The input / output (I / O) interface 705 is also connected to the bus 704.

[0103] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, a mouse, etc.; an output section 707 including, for example, a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc. is installed on the drive 710 as needed so that a computer program read therefrom is installed into the storage section 708 as needed.

[0104] Specifically, according to an embodiment of the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, an embodiment of the present application includes a computer program product including a computer program carried on a computer-readable medium, the computer program including a computer program for performing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network via the communication section 709, and / or installed from the removable medium 711. When the computer program is executed by a central processing unit (CPU) 701, various functions defined in the system of the present application are executed.

[0105] It should be noted that the computer-readable medium shown in the embodiments of the present application can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the above two. A computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable computer program. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted by any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0106] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present application. Among them, each block in the flowchart or block diagram can represent a module, a program segment, or a part of code, and the above module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram or flowchart, and the combination of blocks in the block diagram or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0107] The units involved in the embodiments described in this application can be implemented in software or in hardware, and the described units can also be provided in a processor. Among them, the names of these units do not, in some cases, constitute a limitation to the unit itself.

[0108] Another aspect of this application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor of the computer, the computer is caused to execute the stable isotope mixing model performance evaluation method as described above. The computer-readable storage medium can be included in the electronic device described in the above embodiments, or can exist alone without being assembled into the electronic device.

[0109] Another aspect of this application also provides a computer program product or a computer program. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the stable isotope mixing model performance evaluation method provided in the above various embodiments.

[0110] The above embodiments are only used to exemplarily illustrate the principles and effects of the present invention, rather than to limit the present invention. Any person familiar with this technology can modify or change the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or changes completed by those with ordinary knowledge in the technical field without departing from the spirit and technical ideas disclosed by the present invention should still be covered by the claims of the present invention.

Claims

1. A method for evaluating the performance of a stable isotope mixing model, characterized in that, The method includes: Collect and establish a pollutant source information database, which contains various types of indicators for distinguishing different pollution sources; Based on the pollutant source information database, determine the standard mixing ratio of each pollution source, and create at least two different prior information sets based on the standard mixing ratio; Obtain various types of indicators in the corresponding mixture according to the standard mixing ratio of each pollution source, construct multiple data distribution models with various types of indicators respectively with the various types of indicators of the mixture as expectations or intermediate values, and generate multiple data sets with different sample sizes for each data distribution model; After combining the generated data sets with the corresponding prior information and inputting them into the stable isotope mixing model to be evaluated for operation, record the test results of each stable isotope mixing model to be evaluated; Compare the test results with the standard mixing ratio to obtain the test accuracy of the stable isotope mixing model, so as to evaluate the performance of the used stable isotope mixing model.

2. The performance evaluation method of the stable isotope mixing model according to claim 1, wherein After combining the generated data sets with the corresponding prior information and inputting them into the stable isotope mixing model to be evaluated for operation, it further includes: Define the combination of any data set and prior information as a test scenario; Record the running time of each stable isotope mixing model in different scenarios respectively, and calculate the average running time of multiple stable isotope mixing models in the same scenario; Compare the running time of any stable isotope mixing model with the average running time to obtain the working efficiency of each stable isotope mixing model in different scenarios.

3. The performance evaluation method of the stable isotope mixing model according to claim 2, characterized in that Evaluating the performance of the used stable isotope mixing model includes: Test the test accuracy and working efficiency of different stable isotope mixing models in the same test scenario respectively, or test the test accuracy and working efficiency of the same stable isotope mixing model in different scenarios; Based on the preset conversion rules, convert the test accuracy and the working efficiency into accuracy scores and efficiency scores respectively, and perform weighted calculation on the accuracy scores based on the preset weight ratio to obtain the performance scores of each stable isotope mixing model in the test scenario; Traverse all test scenarios to obtain the performance scores of each stable isotope mixing model in different test scenarios, so as to evaluate the performance of each model.

4. The method for evaluating the performance of the stable isotope mixing model according to claim 1, characterized in that, Comparing the test results with the standard mixing ratio to obtain the test accuracy of the stable isotope mixing model includes: Based on the standard mixing ratio, determine the standard contribution ratio of each pollution source, obtain various types of indicators in the corresponding mixture according to the standard contribution ratio of the pollution source, construct multiple data sets with different sample sizes of data distributions with various types of indicators of the mixture as expectations or intermediate values respectively, and based on the test results of the stable isotope mixing model, obtain the test contribution ratio of each pollution source; Calculate the difference between the standard contribution ratio and the test contribution ratio corresponding to the same pollution source to obtain the test error of the used stable isotope mixing model for each pollution source; Comprehensively calculate the test errors of different pollution sources to obtain the test accuracy of the used stable isotope mixing model.

5. The performance evaluation method of the stable isotope mixing model according to claim 4, wherein Comprehensively calculate the test errors of different pollution sources, including at least one of the following: Calculate the corresponding root mean square error based on the test errors of each pollution source; Calculate the corresponding mean absolute percentage error based on the test errors of each pollution source.

6. The method for evaluating the performance of the stable isotope mixing model according to claim 1, characterized in that, The method further includes: Set the number of pollution sources, the types of tracers, and the number of tracers; Randomly combine the number of pollution sources, the types of tracers, and the number of tracers, and determine each combination result as a test condition to generate multiple test scenarios under different test conditions.

7. The method for evaluating the performance of a stable isotope mixing model according to any one of claims 1-6, wherein The prior information set at least includes a standard information set and an interference information set. Among them, the data mixing ratio of the standard information set satisfies the standard mixing ratio, and the data mixing ratio of the interference information set deviates from the standard mixing ratio; The data distribution model at least includes a normal distribution, a uniform distribution, and a discrete distribution.

8. An apparatus for evaluating the performance of a stable isotope mixing model, characterized in that, The device includes: An information collection module for collecting and establishing a pollutant source information database, which contains various types of indicators for distinguishing different pollution sources; A test preparation module for determining the standard mixing ratio of each pollution source based on the pollutant source information database, and creating at least two different prior information sets based on the standard mixing ratio; A test data generation module for obtaining various types of indicators in the corresponding mixture according to the standard mixing ratio of each pollution source, constructing multiple data distribution models with various types of indicators as expectations or intermediate values respectively according to the various types of indicators of the mixture, and generating multiple data sets with different sample sizes for each data distribution model; A test implementation module for inputting the generated data set combined with the corresponding prior information into the stable isotope mixing model to be evaluated for operation, and recording the test results of each stable isotope mixing model to be evaluated; A model performance evaluation module for comparing the test results with the standard mixing ratio to obtain the test accuracy of the stable isotope mixing model, so as to evaluate the performance of the used stable isotope mixing model.

9. An electronic device, characterized in that, The electronic device includes: One or more processors; A storage device for storing one or more programs, which, when executed by the one or more processors, cause the electronic device to implement the method for evaluating the performance of a stable isotope mixing model according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, A computer program is stored thereon, which, when executed by a processor of the computer, causes the computer to execute the method for evaluating the performance of a stable isotope mixing model according to any one of claims 1 to 7.

Images