Pollutant monitoring system and method using electrochemical biosensor array

By using data processing and modeling methods based on electrochemical biosensor arrays, environmental interference factors are identified and quantified, improving the accuracy and reliability of water quality monitoring and solving the problem of environmental interference in traditional methods.

CN118604074BActive Publication Date: 2024-11-15SHANGHAI KEZE INTELLIGENCE ENVIRONMENT SCI-TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411074216.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-07
Publication Date
2024-11-15
Estimated Expiration
2044-08-07

AI Technical Summary

Technical Problem

In existing technologies, electrochemical biosensors are affected by environmental factors such as temperature, pH, and salinity, as well as the reactivity of organic pollutants, in water quality monitoring, resulting in inaccurate measurement results.

Method used

By employing an electrochemical biosensor array combined with sample data collection, feature extraction, unsupervised training, and supervised training modules, and through orthogonal measurement experiments and feature dimensionality reduction, interfering factors are identified and quantified, and a supervised model capable of predicting organic pollutant concentrations is trained.

Benefits of technology

It improves the accuracy and reliability of organic pollutant concentration measurement, reduces the subjectivity of manually identifying interference factors, and enhances the model's generalization ability and interpretability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118604074B_ABST
    Figure CN118604074B_ABST
Patent Text Reader

Abstract

The invention discloses a pollutant monitoring system and method using an electrochemical biosensor array, and relates to the technical field of pollutant concentration measurement. A sample data collection module is provided to collect measurement sample data, a feature extraction module is provided to obtain the principal component feature vector of each orthogonal measurement test, an unsupervised training module is provided to perform unsupervised clustering on the principal component feature vector of each orthogonal measurement test to obtain clusters of clustering results, and the clustering result clusters are mined to obtain a list of important interference factors, a supervised training module is provided to train a supervised model with a predicted value of organic pollutant concentration as output, and a model application module is provided to obtain the measurement results of organic pollutants in water quality based on actual measurement data and the supervised model, thereby improving the accuracy and reliability of the organic pollutant concentration measurement using the electrochemical biosensor array.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of pollutant concentration measurement, and in particular to a pollutant monitoring system and method using an electrochemical biosensor array. Background Art

[0002] Water quality monitoring is a key task in the field of environmental protection and public health. In particular, organic pollutants, such as pesticides, industrial chemicals, and drug residues, may have serious impacts on ecosystems and human health. Traditional water quality monitoring methods mainly rely on laboratory analysis, usually including techniques such as gas chromatography, liquid chromatography, and mass spectrometry. Although these methods are highly accurate, they often have limitations such as long analysis time, expensive equipment, complex operation, and the inability to achieve real-time on-site monitoring.

[0003] With the development of technology, electrochemical biosensors have become an ideal choice for water quality monitoring due to their high sensitivity, rapid response and portability. Electrochemical biosensors can achieve efficient detection of specific organic pollutants by combining biorecognition elements (such as enzymes, antibodies, microorganisms, etc.) with electrochemical detection technology. However, single electrochemical biosensors are often limited by selectivity and stability in practical applications. Therefore, the development of sensor array-based systems that combine the advantages of multiple sensors to improve the accuracy and reliability of detection has become a current research hotspot.

[0004] The electrochemical biosensor array system can detect multiple organic pollutants simultaneously by integrating multiple sensors, and use data processing and analysis technology to achieve qualitative and quantitative analysis of pollutants. This system can not only monitor water quality in real time, but also provide dynamic change information of pollutants, providing an important basis for water quality management and pollution control.

[0005] However, in practical applications, environmental factors such as temperature, pH value, salinity, and reactive interference or additive effect interference between different organic pollutants will interfere with the response of biosensors and affect the accuracy of measurement results. Therefore, it is necessary to develop advanced data processing and modeling methods to identify and quantify the impact of these interfering factors, so as to improve the reliability and interpretability of detection results.

[0006] The patent application with publication number CN1598537A discloses an on-site optical measurement device and detection method for multiple organic pollutants in seawater, which is mainly composed of a pulsed xenon lamp, a diffraction grating, a sample cell, an absorption light detector, a fluorescence detector, an optical filter, a sampling pump, a control system, etc. The type of the substance is determined based on the spectrum information recorded from the excitation light at a specific wavelength and the fluorescence generated by the excitation light, and the content of the substance is measured based on the absorption spectrum of the specific wavelength. However, this method fails to solve the problem of interference from multiple environmental factors.

[0007] To this end, the present invention proposes a pollutant monitoring system and method using an electrochemical biosensor array. Summary of the invention

[0008] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention proposes a pollutant monitoring system and method using an electrochemical biosensor array, which improves the accuracy and reliability of measuring the concentration of organic pollutants using the electrochemical biosensor array.

[0009] To achieve the above purpose, a pollutant monitoring system using an electrochemical biosensor array is proposed, including a sample data collection module, a feature extraction module, an unsupervised training module, a supervised training module and a model application module; wherein each module is electrically connected;

[0010] The sample data collection module deploys an electrochemical biosensor array in a test water environment, performs N orthogonal measurement tests, and collects Q groups of measurement sample data; N is the number of selected orthogonal measurement tests; N=Q;

[0011] The electrochemical biosensor array is deployed in the test water environment, N orthogonal measurement tests are performed, and Q groups of measurement sample data are collected, which includes the following steps:

[0012] Step 11: Select various types of electrochemical biosensors, and fix each electrochemical biosensor at different positions of the test water environment to form an electrochemical biosensor array;

[0013] Step 12: Connect each electrochemical biosensor to a data acquisition device via a wireless network or a wired network;

[0014] Step 13: Identify all cross-influencing factors;

[0015] Step 14: Divide the value of each cross-influencing factor into several discrete levels;

[0016] Step 15: Generate an orthogonal test table based on orthogonal test theory and encode the discrete level combinations of all influencing factors;

[0017] Step 16: Perform the orthogonal measurement experiment in the order of the orthogonal measurement experiment table. Under each set of orthogonal measurement experiment conditions, add standard organic pollutants of different types and known concentrations, and record the time series of the response values ​​of the electrochemical reactions of each electrochemical biosensor after each orthogonal measurement experiment is implemented;

[0018] Step 17: During each orthogonal measurement experiment, the values ​​of each cross-influencing factor and the concentration of each standard organic pollutant are combined into a set of experimental feature sets;

[0019] The experimental feature set and the response value time series of each electrochemical biosensor constitute a set of measurement sample data;

[0020] The sample data collection module sends the collected measurement sample data to the feature extraction module;

[0021] The feature extraction module performs feature engineering processing on the measured sample data to obtain the result feature matrix of each standard organic pollutant; performs feature dimension reduction on the result feature matrix to obtain the principal component feature vector of each orthogonal measurement test;

[0022] The method for obtaining the result characteristic matrix of each standard organic pollutant comprises the following steps:

[0023] Step 21: for the response value time series of each electrochemical biosensor, extract all statistical features, and all statistical features of each electrochemical biosensor form a set of measurement feature vectors;

[0024] Step 22: construct a preliminary feature matrix for each standard organic pollutant; the size of the preliminary feature matrix is ; Where i is the number of the standard organic pollutant type, Ni is the number of orthogonal measurement experiments with the addition of the i-th standard organic pollutant; M is the number of electrochemical biosensors, and H is the number of elements of the measurement feature vector; each column in the preliminary feature matrix represents a statistical feature of an electrochemical biosensor in an orthogonal measurement experiment;

[0025] Step 23: For Each Electrochemical Biosensor:

[0026] Based on the experimental feature set of each orthogonal measurement experiment and the mean response value of the response value time series of the electrochemical biosensor, a polynomial regression model is used to fit the polynomial function expression between various cross-influencing factors, the concentration of standard organic pollutants and the mean response value;

[0027] Step 24: adding each monomial in the polynomial function expression of each electrochemical biosensor as a new column to the preliminary feature matrix to obtain a result feature matrix;

[0028] The method of performing feature dimension reduction on the result feature matrix to obtain the principal component feature vector of each orthogonal measurement test is as follows:

[0029] The values ​​of all columns of the corresponding rows in the result feature matrix of each orthogonal measurement test are combined into a set of high-dimensional feature vectors;

[0030] Based on the high-dimensional feature vectors of all orthogonal measurement tests, the principal component analysis method is used to convert the high-dimensional feature expression of the result feature matrix into low-dimensional features, and the low-dimensional feature vector expression of each orthogonal measurement test is output as the principal component feature vector;

[0031] The feature extraction module sends the principal component feature vector of each orthogonal measurement test to the unsupervised training module, and sends the resulting feature matrix to the supervised training module;

[0032] Unsupervised training module: for each standard organic pollutant: perform unsupervised clustering on the principal component feature vector of each orthogonal measurement test to obtain clusters of clustering results, and mine the clusters of clustering results to obtain a list of important interference factors;

[0033] For each standard organic pollutant: performing unsupervised clustering on the principal component feature vector of each orthogonal measurement test to obtain clusters of clustering results, and mining the clusters of clustering results to obtain a list of important interference factors includes the following steps:

[0034] Step 31: For the i-th standard organic pollutant:

[0035] The principal component eigenvectors of each orthogonal measurement test with the standard organic pollutant are combined into a The principal component matrix of ; Li is the number of elements of each principal component eigenvector;

[0036] Step 32: Construct a matrix of size A sample distance matrix, each element of which corresponds to: the Euclidean distance between the principal component eigenvectors of two orthogonal measurement tests of the corresponding row and the corresponding column;

[0037] Step 33: Input the sample distance matrix into the K-Means algorithm to obtain K clustering result clusters; K is a preset number parameter of clustering result clusters;

[0038] Step 34: For all cross-influencing factors of each clustering result cluster:

[0039] Then, the principal component analysis method is used to analyze all orthogonal measurement tests in the cluster result cluster, and all cross-influencing factors whose principal component scores in the cluster result cluster are greater than a preset score threshold are obtained as important interference factors of the corresponding type of standard organic pollutants;

[0040] Remove overlaps and merge the important interference factors of all clustering result clusters to obtain a list of important interference factors of the corresponding types of standard organic pollutants;

[0041] The unsupervised training module sends the list of important interference factors to the supervised training module;

[0042] A supervised training module, based on the result feature matrix of each standard organic pollutant and the list of important interference factors, trains a supervised model with the predicted value of organic pollutant concentration as output;

[0043] The method of training the supervised model with the predicted value of organic pollutant concentration as output based on the result feature matrix of each standard organic pollutant and the list of important interference factors is as follows:

[0044] For each criteria organic pollutant:

[0045] From all the columns of the result feature matrix of the standard organic pollutant, all the columns of the initial feature matrix are deleted, and the remaining matrix is ​​used as the target feature matrix;

[0046] For any column item of the target feature matrix, if any cross-influencing factor contained in the column item belongs to the list of important interference factors, then the column item is retained, otherwise, the column item is deleted; the deleted target feature matrix is ​​used as the retained feature matrix;

[0047] Concatenate the retained feature matrix and the initial feature matrix to obtain the supervised feature matrix;

[0048] For the i-th standard organic pollutant:

[0049] Each row in the corresponding supervised feature matrix is ​​used as a training input feature vector, and the concentration of the standard organic pollutant added in the orthogonal measurement test corresponding to each row is used as a concentration label;

[0050] Each set of training input feature vectors is used as the input of the supervised model, and the supervised model takes the predicted value of the pollutant concentration as the output; the supervised model takes the concentration label of each set of training input feature vectors as the prediction target, takes the difference between the predicted value of the pollutant concentration and the concentration label as the prediction error, and takes minimizing the sum of squares of all prediction errors for the concentration label as the training target; the supervised model is trained until the sum of squares of the prediction errors reaches convergence, and the training is stopped; the supervised model is any one of the regression models;

[0051] The supervised training module sends the trained supervised model to the model application module;

[0052] The model application module uses an electrochemical biosensor array to collect actual measurement data of the water body to be monitored, and obtains the measurement results of organic pollutants in water quality based on the actual measurement data and the supervision model;

[0053] The method of obtaining the measurement results of organic pollutants in water quality based on actual measurement data and supervision model is as follows:

[0054] For each criteria organic pollutant:

[0055] Collect the values ​​of the corresponding important interference factors in the water body to be monitored;

[0056] According to the time-ordered sequences of actual response values ​​collected by each electrochemical biosensor array, the statistical features of each sequence are calculated and a set of preliminary feature vectors are formed;

[0057] According to the values ​​of important interference factors in the water body to be monitored, the corresponding values ​​of each column item in the retained feature matrix are calculated, and the values ​​of all column items constitute the influencing factor vector;

[0058] The preliminary feature vector and the influencing factor vector form a set of actual input vectors;

[0059] The actual input vector is input into the supervision model of the corresponding type of standard organic pollutants to obtain the measured value of the organic pollutant concentration output by the supervision model.

[0060] A pollutant monitoring method using an electrochemical biosensor array is proposed, comprising the following steps:

[0061] Step 1: deploy an electrochemical biosensor array in a test water environment, conduct N orthogonal measurement tests, and collect Q groups of measurement sample data;

[0062] Step 2: Perform feature engineering on the measured sample data to obtain the result feature matrix of each standard organic pollutant; perform feature dimension reduction on the result feature matrix to obtain the principal component feature vector of each orthogonal measurement test;

[0063] Step 3: For each standard organic pollutant: perform unsupervised clustering on the principal component feature vector of each orthogonal measurement test to obtain clusters of clustering results, and mine the clusters of clustering results to obtain a list of important interference factors;

[0064] Step 4: Based on the result feature matrix of each orthogonal measurement test and the list of important interference factors, a supervised model with the predicted value of organic pollutant concentration as output is trained;

[0065] Step 5: Use the electrochemical biosensor array to collect actual measurement data of the water body to be monitored, and obtain the measurement results of organic pollutants in water quality based on the actual measurement data and the supervision model.

[0066] An electronic device is proposed, comprising: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0067] The processor executes the above-mentioned pollutant monitoring method using the electrochemical biosensor array by calling the computer program stored in the memory.

[0068] A computer-readable storage medium is provided, on which a rewritable computer program is stored;

[0069] When the computer program is executed on a computer device, the computer device is caused to execute the above-mentioned pollutant monitoring method using an electrochemical biosensor array.

[0070] Compared with the prior art, the present invention has the following beneficial effects:

[0071] The present invention first deploys an electrochemical biosensor array to collect measurement data under different environmental conditions. Then, through techniques such as feature engineering and unsupervised clustering, the inherent patterns and structures of the sample data are automatically discovered, and the key interference factors that have a significant impact on the measurement results are identified. Next, these interference factors are combined with the original measurement data to train a supervised model that can predict the concentration of organic pollutants. Finally, the model is used to analyze the actual collected water data to obtain accurate pollutant concentration measurement results. It effectively combines data-driven and knowledge-driven paradigms, automatically discovers key factors that affect measurements, and incorporates them into the prediction model, thereby improving the generalization ability and interpretability of the model. Compared with traditional methods, this solution avoids the subjectivity of manual identification of interference factors and improves the accuracy and reliability of detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 This is a module connection relationship diagram of a pollutant monitoring system using an electrochemical biosensor array in Example 1 of the present invention;

[0073] Figure 2 is a flow chart of a pollutant monitoring method using an electrochemical biosensor array in Example 2 of the present invention;

[0074] Figure 3 This is a schematic diagram of the structure of an electronic device in Embodiment 3 of the present invention;

[0075] Figure 4 This is a schematic diagram of the computer-readable storage medium structure in Example 4 of the present invention. DETAILED DESCRIPTION

[0076] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0077] Example 1

[0078] like Figure 1As shown, a pollutant monitoring system using an electrochemical biosensor array includes a sample data collection module, a feature extraction module, an unsupervised training module, a supervised training module, and a model application module; wherein each module is electrically connected;

[0079] The sample data collection module deploys an electrochemical biosensor array in a test water environment, performs N orthogonal measurement tests, and collects Q groups of measurement sample data; N is the number of selected orthogonal measurement tests; Q is the number of measurement sample data, N=Q;

[0080] The arrangement of the test water environment is as follows:

[0081] According to the actual situation, select a pool or tank of a certain volume as the simulated water environment;

[0082] Inject clean water into the simulated environment and adjust to the preset baseline water quality parameters such as initial salinity and pH value;

[0083] Install a water circulation system to simulate the water environment to ensure that the water remains evenly mixed; the simulated water environment is used as the test water environment;

[0084] Furthermore, the electrochemical biosensor array is deployed in the test water environment, N orthogonal measurement tests are performed, and Q groups of measurement sample data are collected, including the following steps:

[0085] Step 11: Select various types of electrochemical biosensors, and fix each electrochemical biosensor at different positions in the test water environment to form an electrochemical biosensor array; specifically, the types of the electrochemical biosensors include but are not limited to sensors for recognizing organic substances such as enzymes, antibodies, and microorganisms;

[0086] Step 12: Connect each electrochemical biosensor to a data acquisition device via a wireless network or a wired network, and the electrochemical biosensor sends a response value of the electrochemical reaction to the data acquisition device in real time;

[0087] Step 13: Determine all cross-influencing factors; the cross-influencing factors include but are not limited to temperature, pH value, salinity, dissolved oxygen, etc., which are determined according to the specific physical properties of water quality;

[0088] Step 14: Divide the value of each cross-influencing factor into a number of discrete levels; the number of discrete levels of each cross-influencing factor and the value range corresponding to the discrete levels are determined based on actual experience;

[0089] Step 15: Generate an orthogonal test table based on orthogonal test theory and encode the discrete level combinations of all influencing factors;

[0090] Specifically, Table 1 is an example of the orthogonal test table;

[0091] Table 1 Orthogonal test table

[0092]

[0093] In Table 1 , each row represents an orthogonal measurement experiment, and each column represents the corresponding discrete level;

[0094] Step 16: performing an orthogonal measurement experiment in the order of the orthogonal test table, adding standard organic pollutants of different types and known concentrations under each set of orthogonal measurement experimental conditions, and recording the time series of the response values ​​of the electrochemical reactions of each electrochemical biosensor after each orthogonal measurement experiment is implemented; specifically, the added types and concentrations of the standard organic pollutants are set and recorded by the experimenter; each standard organic pollutant corresponds to an electrochemical biosensor;

[0095] Step 17: During each orthogonal measurement experiment, the values ​​of each cross-influencing factor and the concentration of each standard organic pollutant are combined into a set of experimental feature sets;

[0096] The experimental feature set and the response value time series of each electrochemical biosensor constitute a set of measurement sample data;

[0097] The sample data collection module sends the collected measurement sample data to the feature extraction module;

[0098] The feature extraction module performs feature engineering processing on the measured sample data to obtain the result feature matrix of each standard organic pollutant; performs feature dimension reduction on the result feature matrix to obtain the principal component feature vector of each orthogonal measurement test;

[0099] The feature engineering process of the measured sample data to obtain the result feature matrix of each standard organic pollutant comprises the following steps:

[0100] Step 21: extracting all statistical features from the response value time series of each electrochemical biosensor, wherein all statistical features of each electrochemical biosensor form a set of measurement feature vectors; the statistical features include peak value, potential, curve integral area, etc.;

[0101] Step 22: construct a preliminary feature matrix for each standard organic pollutant; the size of the preliminary feature matrix is ; Where i is the number of the standard organic pollutant type, Ni is the number of orthogonal measurement experiments with the addition of the i-th standard organic pollutant; M is the number of electrochemical biosensors, and H is the number of elements of the measurement feature vector; each column in the preliminary feature matrix represents a statistical feature of an electrochemical biosensor in an orthogonal measurement experiment;

[0102] Step 23: For Each Electrochemical Biosensor:

[0103] Based on the experimental feature set of each orthogonal measurement experiment and the response value mean of the response value time series of the electrochemical biosensor, a polynomial regression model is used to fit the polynomial function expression between various cross-influencing factors, the concentration of standard organic pollutants and the response value mean; a feasible polynomial function expression includes the linear term of the concentration of the standard organic pollutant, the square value of the temperature, the product of the temperature and the pH, etc.;

[0104] Step 24: adding each monomial in the polynomial function expression of each electrochemical biosensor as a new column to the preliminary feature matrix to obtain a result feature matrix; for example, adding the product of temperature and pH in the polynomial function expression of each electrochemical biosensor as a new column to the preliminary feature matrix, so that the number of columns added to the preliminary feature matrix by each electrochemical biosensor is the number of terms in its corresponding polynomial function expression;

[0105] Furthermore, the method of performing feature dimension reduction on the result feature matrix to obtain the principal component feature vector of each orthogonal measurement test is as follows:

[0106] The values ​​of all columns of the corresponding rows in the result feature matrix of each orthogonal measurement test are combined into a set of high-dimensional feature vectors;

[0107] Based on the high-dimensional feature vectors of all orthogonal measurement tests, the principal component analysis method is used to convert the high-dimensional feature expression of the result feature matrix into low-dimensional features, and the low-dimensional feature vector expression of each orthogonal measurement test is output as the principal component feature vector;

[0108] It can be understood that each element in the low-dimensional feature vector constructed by the principal component analysis method (PCA) represents a principal component score. The principal component score is the projection component of the original high-dimensional feature in the principal component space. It retains most of the effective information in the original feature, but has a lower dimension and better noise resistance.

[0109] The feature extraction module sends the principal component feature vector of each orthogonal measurement test to the unsupervised training module, and sends the resulting feature matrix to the supervised training module;

[0110] Unsupervised training module: for each standard organic pollutant: perform unsupervised clustering on the principal component feature vector of each orthogonal measurement test to obtain clusters of clustering results, and mine the clusters of clustering results to obtain a list of important interference factors;

[0111] For each standard organic pollutant: performing unsupervised clustering on the principal component feature vector of each orthogonal measurement test to obtain clusters of clustering results, and mining the clusters of clustering results to obtain a list of important interference factors includes the following steps:

[0112] Step 31: For the i-th standard organic pollutant:

[0113] The principal component eigenvectors of each orthogonal measurement test with the standard organic pollutant are combined into a The principal component matrix of ; Li is the number of elements of each principal component eigenvector;

[0114] Step 32: Construct a matrix of size A sample distance matrix, each element of which corresponds to: the Euclidean distance between the principal component eigenvectors of two orthogonal measurement tests of the corresponding row and the corresponding column;

[0115] Step 33: Input the sample distance matrix into the K-Means algorithm to obtain K clustering result clusters; K is a preset number parameter of clustering result clusters;

[0116] It can be understood that each clustering result cluster contains several orthogonal measurement experiments, and the orthogonal measurement experiments in each cluster have similar environments and measurement properties;

[0117] Step 34: For all cross-influencing factors of each clustering result cluster:

[0118] Then, the principal component analysis method is used to analyze all orthogonal measurement tests in the cluster result cluster, and all cross-influencing factors whose principal component scores in the cluster result cluster are greater than a preset score threshold are obtained as important interference factors of the corresponding type of standard organic pollutants;

[0119] Remove overlaps and merge the important interference factors of all clustering result clusters to obtain a list of important interference factors of the corresponding types of standard organic pollutants;

[0120] The unsupervised training module sends the list of important interference factors to the supervised training module;

[0121] A supervised training module, based on the result feature matrix of each standard organic pollutant and the list of important interference factors, trains a supervised model with the predicted value of organic pollutant concentration as output;

[0122] The method of training the supervised model with the predicted value of organic pollutant concentration as output based on the result feature matrix of each standard organic pollutant and the list of important interference factors is as follows:

[0123] For each criteria organic pollutant:

[0124] From all the columns of the result characteristic matrix of the standard organic pollutant, all the columns of the initial characteristic matrix are deleted, and the remaining matrix is ​​used as the target characteristic matrix; it can be understood that the column items in the target characteristic matrix are all monomial items of various cross-influencing factors;

[0125] For any column item of the target feature matrix, if any cross-influencing factor contained in the column item belongs to the list of important interference factors, then the column item is retained, otherwise, the column item is deleted; the deleted target feature matrix is ​​used as the retained feature matrix;

[0126] Concatenate the retained feature matrix and the initial feature matrix to obtain the supervised feature matrix;

[0127] For the i-th standard organic pollutant:

[0128] Each row in the corresponding supervised feature matrix is ​​used as a training input feature vector, and the concentration of the standard organic pollutant added in the orthogonal measurement test corresponding to each row is used as a concentration label;

[0129] Each set of training input feature vectors is used as the input of the supervised model, and the supervised model takes the predicted value of the pollutant concentration as the output; the supervised model takes the concentration label of each set of training input feature vectors as the prediction target, takes the difference between the predicted value of the pollutant concentration and the concentration label as the prediction error, and takes minimizing the sum of squares of all prediction errors for the concentration label as the training target; the supervised model is trained until the sum of squares of the prediction errors reaches convergence, and the training is stopped; the supervised model is any one of the regression models, and the regression model includes but is not limited to the SVR model or the polynomial regression model, etc.;

[0130] The supervised training module sends the trained supervised model to the model application module;

[0131] The model application module uses an electrochemical biosensor array to collect actual measurement data of the water body to be monitored, and obtains the measurement results of organic pollutants in water quality based on the actual measurement data and the supervision model;

[0132] The method of obtaining the measurement results of organic pollutants in water quality based on actual measurement data and supervision model is as follows:

[0133] For each criteria organic pollutant:

[0134] Collect the values ​​of the corresponding important interference factors in the water body to be monitored;

[0135] According to the time-ordered sequence of actual response values ​​collected by each electrochemical biosensor array, the statistical features of each sequence are calculated and a set of preliminary feature vectors are formed; it can be understood that the form of the preliminary feature vector is consistent with the form of each row in the preliminary feature matrix;

[0136] According to the values ​​of important interference factors in the water body to be monitored, the corresponding values ​​of each column item in the retained feature matrix are calculated, and the values ​​of all column items constitute the influencing factor vector;

[0137] The preliminary feature vector and the influencing factor vector form a set of actual input vectors;

[0138] The actual input vector is input into the supervision model of the corresponding type of standard organic pollutants to obtain the measured value of the organic pollutant concentration output by the supervision model.

[0139] Example 2

[0140] like Figure 2 As shown, the pollutant monitoring method using an electrochemical biosensor array comprises the following steps:

[0141] Step 1: deploy an electrochemical biosensor array in a test water environment, conduct N orthogonal measurement tests, and collect Q groups of measurement sample data;

[0142] Step 2: Perform feature engineering on the measured sample data to obtain the result feature matrix of each standard organic pollutant; perform feature dimension reduction on the result feature matrix to obtain the principal component feature vector of each orthogonal measurement test;

[0143] Step 3: For each standard organic pollutant: perform unsupervised clustering on the principal component feature vector of each orthogonal measurement test to obtain clusters of clustering results, and mine the clusters of clustering results to obtain a list of important interference factors;

[0144] Step 4: Based on the result feature matrix of each orthogonal measurement test and the list of important interference factors, a supervised model with the predicted value of organic pollutant concentration as output is trained;

[0145] Step 5: Use the electrochemical biosensor array to collect actual measurement data of the water body to be monitored, and obtain the measurement results of organic pollutants in water quality based on the actual measurement data and the supervision model.

[0146] Example 3

[0147] Figure 3 1 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Figure 3As shown, the electronic device 100 may include one or more processors and one or more memories. The memories store computer readable codes, which, when executed by the one or more processors, may execute the pollutant monitoring method using the electrochemical biosensor array as described above.

[0148] The method or device according to the embodiment of the present application can also be used by Figure 3 The electronic device architecture shown in FIG. Figure 3 As shown, the electronic device 100 may include a bus 101, one or more CPUs 102, a ROM 103, a RAM 104, a communication port 105 connected to a network, an input / output component 106, a hard disk 107, etc. The storage device in the electronic device 100, such as the ROM 103 or the hard disk 107, may store the pollutant monitoring method using the electrochemical biosensor array provided in the present application.

[0149] Furthermore, the electronic device 100 may further include a user interface 108. Figure 3 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 3 One or more components of an electronic device are shown.

[0150] Example 4

[0151] Figure 4 Schematic diagram of a computer-readable storage medium structure provided by an embodiment of the present application. Figure 4 As shown, computer-readable instructions are stored on the computer-readable storage medium 200. When the computer-readable instructions are executed by the processor, the pollutant monitoring method using the electrochemical biosensor array according to the embodiment of the present application described with reference to the above figures can be executed. The computer-readable storage medium 200 includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0152] In addition, according to the implementation of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, the non-transitory machine-readable storage medium stores machine-readable instructions, the machine-readable instructions can be run by a processor to execute instructions corresponding to the method steps provided by the present application, and when the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0153] The methods, apparatuses, and devices of the present application may be implemented in many ways. For example, the methods, apparatuses, and devices of the present application may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order of steps for the method is for illustration only, and the steps of the method of the present application are not limited to the order specifically described above unless otherwise specified. In addition, in some embodiments, the present application may also be implemented as programs recorded in a recording medium, which include machine-readable instructions for implementing the method according to the present application. Therefore, the present application also covers recording media storing programs for executing the method according to the present application.

[0154] In addition, the parts of the above-mentioned technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.

[0155] The specific implementation modes as described above further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation mode of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0156] The above preset parameters or preset thresholds are all set by technicians in this field according to actual conditions or obtained through large-scale data simulation.

[0157] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A pollutant monitoring system using an electrochemical biosensor array, characterized in that: It includes a sample data collection module, a feature extraction module, an unsupervised training module, a supervised training module and a model application module; wherein each module is electrically connected; The sample data collection module deploys an electrochemical biosensor array in the test water environment, performs N orthogonal measurement tests, collects Q groups of measurement sample data; N is the number of selected orthogonal measurement tests, and sends the collected measurement sample data to the feature extraction module; N=Q; The feature extraction module performs feature engineering processing on the measurement sample data to obtain the result feature matrix of each standard organic pollutant; performs feature dimension reduction on the result feature matrix to obtain the principal component feature vector of each orthogonal measurement test; and sends the principal component feature vector of each orthogonal measurement test to the unsupervised training module, and sends the result feature matrix to the supervised training module; Unsupervised training module, for each standard organic pollutant: perform unsupervised clustering on the principal component feature vector of each orthogonal measurement test to obtain clusters of clustering results, and mine the clustering result clusters to obtain a list of important interference factors, and send the list of important interference factors to the supervised training module; For each standard organic pollutant: performing unsupervised clustering on the principal component feature vector of each orthogonal measurement test to obtain clusters of clustering results, and mining the clusters of clustering results to obtain a list of important interference factors includes the following steps: Step 31: For the i-th standard organic pollutant: The principal component eigenvectors of each orthogonal measurement test with the standard organic pollutant are combined into a The principal component matrix of ; Li is the number of elements of each principal component eigenvector; Step 32: Construct a matrix of size A sample distance matrix, each element of which corresponds to: the Euclidean distance between the principal component eigenvectors of two orthogonal measurement tests of the corresponding row and the corresponding column; Step 33: Input the sample distance matrix into the K-Means algorithm to obtain K clustering result clusters; K is a preset number parameter of clustering result clusters; Step 34: For all cross-influencing factors of each clustering result cluster; The supervised training module trains a supervised model with the predicted value of organic pollutant concentration as output based on the result feature matrix of each standard organic pollutant and the list of important interference factors, and sends the trained supervised model to the model application module; The model application module uses an electrochemical biosensor array to collect actual measurement data of the water body to be monitored, and obtains the measurement results of organic pollutants in water quality based on the actual measurement data and the supervision model.

2. The pollutant monitoring system using an electrochemical biosensor array according to claim 1, characterized in that: The electrochemical biosensor array is deployed in the test water environment, N orthogonal measurement tests are performed, and Q groups of measurement sample data are collected, which includes the following steps: Step 11: Select various types of electrochemical biosensors, and fix each electrochemical biosensor at different positions of the test water environment to form an electrochemical biosensor array; Step 12: Connect each electrochemical biosensor to a data acquisition device via a wireless network or a wired network; Step 13: Identify all cross-influencing factors; Step 14: Divide the value of each cross-influencing factor into discrete levels; Step 15: Generate an orthogonal test table based on orthogonal test theory and encode the discrete level combinations of all cross-influencing factors; Step 16: Perform the orthogonal measurement experiment in the order of the orthogonal measurement experiment table. Under each set of orthogonal measurement experiment conditions, add standard organic pollutants of different types and known concentrations, and record the time series of the response values ​​of the electrochemical reactions of each electrochemical biosensor after each orthogonal measurement experiment is implemented; Step 17: During each orthogonal measurement experiment, the values ​​of each cross-influencing factor and the concentration of each standard organic pollutant are combined into a set of experimental feature sets; The experimental feature set and the response value time series of each electrochemical biosensor constitute a set of measurement sample data.

3. The pollutant monitoring system using an electrochemical biosensor array according to claim 2, characterized in that: The method for obtaining the result characteristic matrix of each standard organic pollutant comprises the following steps: Step 21: for the response value time series of each electrochemical biosensor, extract all statistical features, and all statistical features of each electrochemical biosensor form a set of measurement feature vectors; Step 22: construct a preliminary feature matrix for each standard organic pollutant; the size of the preliminary feature matrix is ; Where i is the number of the standard organic pollutant type, Ni is the number of orthogonal measurement experiments with the addition of the i-th standard organic pollutant; M is the number of electrochemical biosensors, and H is the number of elements of the measurement feature vector; each column in the preliminary feature matrix represents a statistical feature of an electrochemical biosensor in an orthogonal measurement experiment; Step 23: For Each Electrochemical Biosensor: Based on the experimental feature set of each orthogonal measurement experiment and the mean response value of the response value time series of the electrochemical biosensor, a polynomial regression model is used to fit the polynomial function expression between various cross-influencing factors, the concentration of standard organic pollutants and the mean response value; Step 24: Add each monomial in the polynomial function expression of each electrochemical biosensor as a new column to the preliminary feature matrix to obtain a result feature matrix.

4. The pollutant monitoring system using an electrochemical biosensor array according to claim 3, characterized in that: The method of performing feature dimension reduction on the result feature matrix to obtain the principal component feature vector of each orthogonal measurement test is as follows: The values ​​of all columns of the corresponding rows in the result feature matrix of each orthogonal measurement test are combined into a set of high-dimensional feature vectors; Based on the high-dimensional feature vectors of all orthogonal measurement tests, the principal component analysis method is used to convert the high-dimensional feature expression of the result feature matrix into low-dimensional features, and the low-dimensional feature vector expression of each orthogonal measurement test is output as the principal component feature vector.

5. The pollutant monitoring system using an electrochemical biosensor array according to claim 4, characterized in that: For each standard organic pollutant: performing unsupervised clustering on the principal component feature vector of each orthogonal measurement test to obtain a clustering result cluster includes the following steps: Step 31: For the i-th standard organic pollutant: The principal component eigenvectors of each orthogonal measurement test with the standard organic pollutant are combined into a The principal component matrix of ; Li is the number of elements of each principal component eigenvector; Step 32: Construct a matrix of size A sample distance matrix, each element of which corresponds to: the Euclidean distance between the principal component eigenvectors of two orthogonal measurement tests of the corresponding row and the corresponding column; Step 33: Input the sample distance matrix into the K-Means algorithm to obtain K clustering result clusters; K is a preset parameter for the number of clustering result clusters.

6. The pollutant monitoring system using an electrochemical biosensor array according to claim 5, characterized in that: The method of mining the clustering result clusters to obtain a list of important interference factors is as follows: For all cross-influencing factors of each clustering result cluster: Then, the principal component analysis method is used to analyze all orthogonal measurement tests in the cluster result cluster, and all cross-influencing factors whose principal component scores in the cluster result cluster are greater than a preset score threshold are obtained as important interference factors of the corresponding type of standard organic pollutants; The important interference factors of all clustering result clusters were removed and merged to obtain a list of important interference factors of the corresponding types of standard organic pollutants.

7. The pollutant monitoring system using an electrochemical biosensor array according to claim 6, characterized in that: The method of training the supervised model with the predicted value of organic pollutant concentration as output based on the result feature matrix of each standard organic pollutant and the list of important interference factors is as follows: For each criteria organic pollutant: From all the columns of the result feature matrix of the standard organic pollutant, all the columns of the initial feature matrix are deleted, and the remaining matrix is ​​used as the target feature matrix; For any column item of the target feature matrix, if any cross-influencing factor contained in the column item belongs to the list of important interference factors, then the column item is retained, otherwise, the column item is deleted; the deleted target feature matrix is ​​used as the retained feature matrix; Concatenate the retained feature matrix and the initial feature matrix to obtain the supervised feature matrix; For the i-th standard organic pollutant: Each row in the corresponding supervised feature matrix is ​​used as a training input feature vector, and the concentration of the standard organic pollutant added in the orthogonal measurement test corresponding to each row is used as a concentration label; Each set of training input feature vectors is used as the input of a supervised model, and the supervised model takes the predicted value of the pollutant concentration as the output; the supervised model takes the concentration label of each set of training input feature vectors as the prediction target, and the difference between the predicted value of the pollutant concentration and the concentration label as the prediction error, and takes minimizing the sum of squares of all prediction errors for the concentration label as the training target; the supervised model is trained until the sum of squares of the prediction errors converges, and the training is stopped; the supervised model is any one of the regression models.

8. The pollutant monitoring system using an electrochemical biosensor array according to claim 7, characterized in that: The method of obtaining the measurement results of organic pollutants in water quality based on actual measurement data and supervision model is as follows: For each criteria organic pollutant: Collect the values ​​of the corresponding important interference factors in the water body to be monitored; According to the time-ordered sequences of actual response values ​​collected by the electrochemical biosensor array, the statistical features of each sequence are calculated and a set of preliminary feature vectors are formed; According to the values ​​of important interference factors in the water body to be monitored, the corresponding values ​​of each column item in the retained feature matrix are calculated, and the values ​​of all column items constitute the influencing factor vector; The preliminary feature vector and the influencing factor vector form a set of actual input vectors; The actual input vector is input into the supervision model of the corresponding type of standard organic pollutants to obtain the measured value of the organic pollutant concentration output by the supervision model.

9. A pollutant monitoring method using an electrochemical biosensor array, which is implemented based on a pollutant monitoring system using an electrochemical biosensor array as claimed in any one of claims 1 to 8, characterized in that: The following steps are involved: Step 1: deploy an electrochemical biosensor array in the test water environment, conduct N orthogonal measurement tests, and collect Q groups of measurement sample data; N=Q; Step 2: Perform feature engineering on the measured sample data to obtain the result feature matrix of each standard organic pollutant; perform feature dimension reduction on the result feature matrix to obtain the principal component feature vector of each orthogonal measurement test; Step 3: For each standard organic pollutant: perform unsupervised clustering on the principal component feature vector of each orthogonal measurement test to obtain clusters of clustering results, and mine the clusters of clustering results to obtain a list of important interference factors; Step 4: Based on the result feature matrix of each orthogonal measurement test and the list of important interference factors, a supervised model with the predicted value of organic pollutant concentration as output is trained; Step 5: Use the electrochemical biosensor array to collect actual measurement data of the water body to be monitored, and obtain the measurement results of organic pollutants in water quality based on the actual measurement data and the supervision model.

10. An electronic device, characterized in that: include: processor and memory, wherein The memory stores a computer program that can be called by the processor; The processor executes the pollutant monitoring method using an electrochemical biosensor array as claimed in claim 9 in the background by calling the computer program stored in the memory.

11. A computer-readable storage medium, characterized in that: A rewritable computer program is stored thereon; When the computer program is executed on a computer device, the computer device is enabled to execute the pollutant monitoring method using an electrochemical biosensor array as claimed in claim 9 in the background.