Method and system for detecting various antibiotics in water body based on artificial intelligence

Abstract

The invention discloses a method and a system for detecting various antibiotics in a water body based on artificial intelligence, a water body to be detected is treated by a solid phase extraction column, and spectrum data is detected by combining a high performance liquid chromatography tandem mass spectrometer; according to the method, multi-dimensional feature data including retention time, parent ion mass-to-charge ratio, fragment ion mass-to-charge ratio, fragment ion strength, peak symmetry parameters and signal-to-noise ratio are extracted, a multi-dimensional feature matrix is constructed, feature information of a target compound can be comprehensively represented, and a rich and reliable basis is provided for subsequent analysis; a pre-trained deep learning model is used for performing multi-target recognition on the multi-dimensional feature matrix, so that qualitative and quantitative analysis of various antibiotics can be realized, and the detection efficiency and the result reliability are remarkably improved; the method has wide applicability and high practical application value, and an efficient and accurate technical means is provided for monitoring and evaluation of antibiotic pollution in the water body.

Description

Technical Field

[0001] This application relates to the field of analytical detection technology, and particularly to a method and system for detecting multiple antibiotics in water based on artificial intelligence. Background Art

[0002] With the rapid development of industrialization and agriculturalization, the problem of antibiotic abuse has become increasingly serious. A large number of antibiotics enter the water environment through wastewater discharge, posing a serious threat to the ecological system and human health. Traditional antibiotic detection methods usually rely on laboratory analysis, such as high-performance liquid chromatography (HPLC) and mass spectrometry (MS) and other technologies. However, these methods involve a large amount of manual data analysis work and are difficult to meet the needs of large-scale water body monitoring. In recent years, the application of artificial intelligence technology in the field of environmental monitoring has gradually increased. However, how to use artificial intelligence to achieve efficient and accurate detection of multiple antibiotics in water is still a technical problem to be solved urgently. Summary of the Invention

[0003] This application provides a method and system for detecting multiple antibiotics in water based on artificial intelligence to solve the technical problem of large amounts of data in the large-scale analysis of existing water body antibiotics.

[0004] To solve the above technical problems, in a first aspect, an embodiment of this application provides a method for detecting multiple antibiotics in water based on artificial intelligence, including: Treating the water to be tested with a solid-phase extraction column and detecting the spectral data of the water to be tested using a high-performance liquid chromatography-tandem mass spectrometer; Based on the spectral data, extracting multi-dimensional feature data including retention time, precursor ion mass-to-charge ratio, fragment ion mass-to-charge ratio, fragment ion intensity, peak symmetry parameter, and signal-to-noise ratio; Based on the multi-dimensional feature data, constructing a multi-dimensional feature matrix; Using a pre-trained deep learning model to perform multi-target recognition on the multi-dimensional feature matrix and output the qualitative analysis results and quantitative analysis results of multiple antibiotics. The deep learning model is trained based on the spectral data of multiple antibiotic standard solution samples and the spectral data set of multiple water body samples. The spectral data set includes spectral data under different concentrations, different matrices, and different interference conditions.

[0005] In some of these embodiments, the treating the water to be tested with a solid-phase extraction column and detecting the spectral data of the water to be tested using a high-performance liquid chromatography-tandem mass spectrometer includes: The water sample to be tested is adsorbed and concentrated by an HLB solid-phase extraction column, eluted with a methanol solution, blown to near dryness with nitrogen at 40 °C, and fixed volume is made with 10% methanol aqueous solution. It is detected by a high performance liquid chromatography tandem mass spectrometer using acetonitrile - 5 mmol / L ammonium acetate aqueous solution and acetonitrile - 10 mmol / L trifluoroacetic acid aqueous solution as the mobile phase to obtain the spectral data of the water sample to be tested.

[0006] In some of these embodiments, based on the spectral data, extracting multi-dimensional feature data including retention time, precursor ion mass-to-charge ratio, fragment ion mass-to-charge ratio, fragment ion intensity, peak symmetry parameter, and signal-to-noise ratio, includes: Performing wavelet transform and time alignment on the spectral data to obtain target spectral data; Using a preset spectral feature extraction model to perform feature extraction on the target spectral data to obtain multi-dimensional feature data including the retention time, precursor ion mass-to-charge ratio, fragment ion mass-to-charge ratio, fragment ion intensity, peak symmetry parameter, and signal-to-noise ratio. The preset spectral feature extraction model is a model trained by a machine learning algorithm based on a preset spectral training set.

[0007] In some of these embodiments, based on the multi-dimensional feature data, constructing a multi-dimensional feature matrix, includes: Calculating the data center and interquartile range of each dimension of feature data in the multi-dimensional feature data; Using an outlier-resistant standardization algorithm to determine the original feature values of each dimension of feature data according to the data center and interquartile range; Based on the original feature values, performing median replacement or truncation processing on the outliers in each dimension of feature data to obtain target multi-dimensional feature data; Combining the target multi-dimensional feature data to construct the multi-dimensional feature matrix.

[0008] In some of these embodiments, using a pre-trained deep learning model to perform multi-target recognition on the multi-dimensional feature matrix and output qualitative analysis results and quantitative analysis results of multiple antibiotics, includes: Inputting the multi-dimensional feature matrix into the deep learning model, outputting prediction results of multiple antibiotic types through the softmax function, and outputting concentration prediction results of multiple antibiotics through the linear regression layer. The deep learning model is a convolutional neural network, a recurrent neural network, a long short-term memory network, a graph neural network, or a multi-task learning model.

[0009] In some of these embodiments, the antibiotics include sulfathiazole, sulfamonomethoxine, sulfamethoxydiazine, sulfafurazole, trimethoprim, sulfacetamide, sulfapyridine, sulfadiazine, sulfamethoxazole, sulfamerazine, sulfamethizole, sulfadimidine, sulfamethoxypyridazine, sulfachloropyridazine, sulfachinoxaline, sulfadimethoxine, sulfadoxine, sulfaphenazole, orbifloxacin, danofloxacin, enrofloxacin, fleroxacin, ciprofloxacin, lomefloxacin, norfloxacin, pefloxacin, sarafloxacin, difloxacin, ofloxacin, enoxacin, chloramphenicol, oxytetracycline, tetracycline, demethylchlortetracycline, norvancomycin, penicillin, cephalosporin, and erythromycin.

[0010] In a second aspect, an embodiment of the present application provides a system for detecting multiple antibiotics in water based on artificial intelligence, including a solid-phase extraction device, an ultra-high performance liquid chromatograph, a triple quadrupole mass spectrometer, and a computer device, where the computer device is used to implement the method for detecting multiple antibiotics in water based on artificial intelligence as described above.

[0011] Compared with the prior art, the present application has at least the following beneficial effects: By treating the water sample to be detected with a solid-phase extraction column and combining the detection spectrum data of a high-performance liquid chromatography tandem mass spectrometer, the target compounds in the water can be effectively enriched and separated, significantly improving the detection sensitivity and accuracy; extracting multi-dimensional characteristic data including retention time, precursor ion mass-to-charge ratio, fragment ion mass-to-charge ratio, fragment ion intensity, peak symmetry parameter, and signal-to-noise ratio, and constructing a multi-dimensional characteristic matrix, which can comprehensively characterize the characteristic information of the target compounds and provide rich and reliable basis for subsequent analysis; using a pre-trained deep learning model to perform multi-target recognition on the multi-dimensional characteristic matrix, the qualitative and quantitative analysis of multiple antibiotics can be realized, significantly improving the detection efficiency and the reliability of the results; this method is applicable to water samples under different concentrations, different matrices, and different interference conditions, has wide applicability and high practical application value, and provides an efficient and accurate technical means for the monitoring and evaluation of antibiotic pollution in water. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 is a schematic flowchart of the method for detecting multiple antibiotics in water based on artificial intelligence shown in an embodiment of the present application; Figure 2 is a structural block diagram of the system for detecting multiple antibiotics in water based on artificial intelligence shown in an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0013] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without making creative efforts belong to the scope of protection of the present application.

[0014] Please refer to Figure 1 , Figure 1 , which is a schematic flowchart of a method for detecting multiple antibiotics in water based on artificial intelligence provided by an embodiment of the present application. The method for detecting multiple antibiotics in water based on artificial intelligence in this embodiment includes steps S101 to S104, which are described in detail as follows: Step S101, treating the water sample to be detected through a solid-phase extraction column, and detecting the spectral data of the water sample to be detected by using a high-performance liquid chromatography-mass spectrometry (HPLC-MS) instrument.

[0015] In this step, treating the water sample to be detected through a solid-phase extraction column can effectively enrich and separate the target compounds, reduce the influence of interfering substances, and significantly improve the sensitivity and selectivity of detection. Combining with the detection of spectral data by a high-performance liquid chromatography-mass spectrometry instrument, high-precision qualitative and quantitative analysis of the target compounds can be realized, which is applicable to the detection of trace pollutants in complex water bodies and provides a reliable technical means for water quality monitoring and pollution assessment.

[0016] In some embodiments, step S101 includes: Adsorbing and concentrating the water sample to be detected through an HLB solid-phase extraction column, eluting with a methanol solution, blowing to near dry with nitrogen at 40 °C, diluting to a constant volume with a 10% methanol aqueous solution, and detecting in a high-performance liquid chromatography-mass spectrometry instrument with acetonitrile-5 mmol / L ammonium acetate aqueous solution and acetonitrile-10 mmol / L trifluoroacetic acid aqueous solution as the mobile phases to obtain the spectral data of the water sample to be detected.

[0017] In this embodiment, the spectral data includes but is not limited to the total ion chromatogram, precursor ion chromatogram, and fragment ion chromatogram. This embodiment can effectively enrich and concentrate the target compounds, reduce the influence of interfering substances, and significantly improve the detection sensitivity and accuracy. By selecting appropriate mobile phases, efficient separation and detection of compounds with different polarities and chemical properties can be achieved, which is applicable to the analysis of trace target substances in complex water bodies. Moreover, this method is simple to operate and has high sensitivity, and can provide reliable technical support for water quality monitoring and pollution assessment.

[0018] Step S102, based on the spectral data, extracting multi-dimensional characteristic data including retention time, precursor ion mass-to-charge ratio, fragment ion mass-to-charge ratio, fragment ion intensity, peak symmetry parameter, and signal-to-noise ratio; In this step, from the spectral data obtained by the high performance liquid chromatography-mass spectrometry (HPLC-MS), the retention time of the target compound is extracted, which is helpful for the qualitative analysis of the compound. The mass spectrometry data is analyzed to extract the mass-to-charge ratio of the parent ion and the mass-to-charge ratio of the fragment ions. These parameters reflect the molecular structure information of the compound. The intensity of the fragment ions is measured to evaluate the relative content of the compound. The peak symmetry parameter and the signal-to-noise ratio are calculated to evaluate the quality of the chromatographic peak and the reliability of the data. By integrating these parameters, a comprehensive data set containing multi-dimensional features is formed, providing comprehensive information support for subsequent analysis.

[0019] The extraction of multi-dimensional feature data in this step can significantly improve the qualitative and quantitative analysis capabilities of the target compound. The combination of the retention time, the mass-to-charge ratio of the parent ion and the mass-to-charge ratio of the fragment ions improves the accuracy of compound identification. The introduction of the fragment ion intensity and the peak symmetry parameter enhances the reliability of the quantitative analysis. The evaluation of the signal-to-noise ratio ensures the quality of the data and reduces the influence of interference factors. Through the comprehensive application of multi-dimensional features, this method can effectively detect the target compound in complex water samples, providing an efficient and accurate technical means for water quality monitoring and pollution assessment.

[0020] In some embodiments, step S102 includes: Performing wavelet transform and time alignment on the spectral data to obtain target spectral data; Using a preset spectral feature extraction model to extract features from the target spectral data, obtaining multi-dimensional feature data including the retention time, the mass-to-charge ratio of the parent ion, the mass-to-charge ratio of the fragment ions, the intensity of the fragment ions, the peak symmetry parameter and the signal-to-noise ratio. The preset spectral feature extraction model is a model trained by a machine learning algorithm based on a preset spectral training set.

[0021] In this embodiment, the mass-to-charge ratio (m / z) of the parent ion is extracted from the mass spectrometry data. The mass-to-charge ratio of the parent ion is a direct reflection of the chemical properties of the compound and is used for subsequent qualitative analysis; the mass-to-charge ratio (m / z) of the fragment ions is extracted from the mass spectrometry data. The mass-to-charge ratio of the fragment ions reflects the fragmentation pattern of the compound and is used to further verify the structure of the compound; the signal intensity of the fragment ions is extracted from the mass spectrometry data. The fragment ion intensity reflects the concentration of the compound and is used for quantitative analysis; by calculating the ratio of the left and right half-peak widths of the chromatographic peak, the peak symmetry parameter is extracted. The peak symmetry parameter is used to evaluate the quality of the chromatographic peak, thereby judging the reliability of the detection result; by calculating the ratio of the signal intensity to the background noise, the signal-to-noise ratio is extracted. The signal-to-noise ratio is used to evaluate the sensitivity of the detection and the reliability of the data.

[0022] Step S103, based on the multi-dimensional feature sequence, construct a multi-dimensional feature matrix.

[0023] In this step, the multi-dimensional features of each water sample to be tested are arranged in a certain order to form a feature sequence. For example, the feature sequence of each sample is [retention time, precursor ion mass-to-charge ratio, fragment ion mass-to-charge ratio, fragment ion intensity, peak symmetry parameter, signal-to-noise ratio], to ensure the consistency and comparability of the features of each sample during data processing. The feature sequences of multiple samples are arranged row by row to form a two-dimensional or multi-dimensional feature matrix. For example, if there are N samples and each sample has M features, the dimension of the feature matrix is N×M. If the feature itself has a multi-dimensional structure (for example, the combination of retention time and precursor ion mass-to-charge ratio), the feature matrix can be extended to a multi-dimensional structure, such as a three-dimensional or four-dimensional array.

[0024] In this step, the multi-dimensional feature matrix can comprehensively reflect the characteristic information of the target compound, including chemical properties and chromatographic behavior, which helps to improve the recognition accuracy and robustness of the model; the rich characteristic information enables the model to better capture the characteristic patterns of the target compound, thereby improving the detection precision.

[0025] In some embodiments, step S103 includes: Calculate the data center and interquartile range of each dimension of feature data in the multi-dimensional feature data; Using the outlier-resistant standardization algorithm, determine the original feature values of each dimension of feature data according to the data center and interquartile range; Based on the original feature values, perform median replacement or truncation processing on the outliers in each dimension of feature data to obtain the target multi-dimensional feature data; Combine the target multi-dimensional feature data to construct the multi-dimensional feature matrix.

[0026] In this embodiment, statistical analysis is performed on each dimension of feature data to calculate its data center (such as the median) and interquartile range (IQR) to reflect the central position and distribution range of the data; using the outlier-resistant standardization algorithm, determine the standardization parameters of each dimension of feature data according to the calculated data center and interquartile range to reduce the influence of outliers on the standardization process; detect and identify the outliers in each dimension of feature data, and use the median replacement or truncation processing method to correct the outliers to ensure the stability and consistency of the data; combine the processed target multi-dimensional feature data to form a structured multi-dimensional feature matrix as the basis for subsequent analysis and modeling.

[0027] In this embodiment, by calculating the data center and interquartile range of each-dimensional feature data, the central position and distribution range of the data can be effectively identified, providing a reliable basis for subsequent standardization and outlier processing. The application of the outlier-resistant standardization algorithm significantly improves the robustness of the data and reduces the influence of extreme values on the standardization process. Median replacement or truncation processing of outliers further enhances the stability and consistency of the data, ensuring the accuracy and reliability of the data. The constructed target multi-dimensional feature matrix provides a high-quality data foundation for subsequent analysis and modeling, can effectively support complex machine learning tasks, and improve the performance and prediction accuracy of the model.

[0028] Step S104: Use a pre-trained deep learning model to perform multi-target recognition on the multi-dimensional feature matrix, and output qualitative analysis results and quantitative analysis results of multiple antibiotics. The deep learning model is trained based on the spectral data of multiple antibiotic standard solution samples and the spectral data sets of multiple water samples. The spectral data sets include spectral data under different concentrations, different matrices, and different interference conditions.

[0029] In this step, using a pre-trained deep learning model to perform multi-target recognition on the multi-dimensional feature matrix can efficiently and accurately output qualitative analysis results and quantitative analysis results of multiple antibiotics. This model is trained based on spectral data sets containing different concentrations, matrices, and interference conditions, and has good adaptability and generalization ability, and can work stably in complex water environments. Through the multi-target recognition ability of the deep learning model, not only can multiple antibiotics be detected simultaneously, but also quantitative information can be provided, providing comprehensive technical support for water quality monitoring and pollution assessment.

[0030] In some embodiments, step S104 includes: Input the multi-dimensional feature matrix into the deep learning model, output prediction results of multiple antibiotic types through the softmax function, and output concentration prediction results of multiple antibiotics through the linear regression layer. The deep learning model is a convolutional neural network, a recurrent neural network, a long short-term memory network, a graph neural network, or a multi-task learning model.

[0031] In this embodiment, when inputting the multi-dimensional feature matrix into the deep learning model, it is first necessary to preprocess and format the feature data to meet the input requirements of the model. The deep learning model (such as convolutional neural network, recurrent neural network, long short-term memory network, graph neural network or multi-task learning model) performs feature extraction and non-linear transformation on the input multi-dimensional feature matrix. For the prediction task of antibiotic types, the model outputs the probability distribution of various antibiotics through the softmax function to achieve the multi-classification goal. For the prediction task of antibiotic concentration, the model outputs continuous concentration values through the linear regression layer to complete the regression task. The whole process makes full use of the multi-task learning ability of the model and can complete the classification and regression tasks simultaneously in the same framework.

[0032] Through the processing of the multi-dimensional feature matrix by the deep learning model in this embodiment, the efficient and accurate identification and concentration prediction of various antibiotics can be achieved. The application of the softmax function ensures the accuracy of antibiotic type prediction, while the introduction of the linear regression layer improves the accuracy of concentration prediction. The multi-task learning ability of the model enables the classification and regression tasks to promote each other and improves the overall performance. In addition, the model is trained based on the spectral data sets of standard solution samples and water body samples of various antibiotics, covering data under different concentrations, matrices and interference conditions, significantly enhancing the generalization ability and robustness of the model. This method provides an efficient and reliable solution for the detection of antibiotics in complex water environments and has important practical application value.

[0033] Optionally, the antibiotics include sulfathiazole, sulfamonomethoxine, sulfamethoxydiazine, sulfisoxazole, trimethoprim, sulfacetamide, sulfapyridine, sulfadiazine, sulfamethoxazole, sulfamerazine, sulfamethizole, sulfadimidine, sulfamethoxypyridazine, sulfachloropyridazine, sulfachinoxaline, sulfadimethoxine, sulfadoxine, sulfaphenazole, orbifloxacin, danofloxacin, enrofloxacin, fleroxacin, ciprofloxacin, lomefloxacin, norfloxacin, pefloxacin, sarafloxacin, difloxacin, ofloxacin, enoxacin, chloramphenicol, oxytetracycline, tetracycline, demethylchlortetracycline, demethylvancomycin, penicillin, cephalosporin and erythromycin.

[0034] As Figure 2 shown, Figure 2 FIG. is a structural block diagram of a system for detecting multiple antibiotics in water based on artificial intelligence provided by an embodiment of the present application, including a solid-phase extraction device 201, an ultra-high performance liquid chromatograph 202, a triple quadrupole mass spectrometer 203 and a computer device 204, and the computer device 204 is used to implement the method for detecting multiple antibiotics in water based on artificial intelligence as described above.

[0035] In several embodiments provided by the present application, it can be understood that each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the module, the program segment, or the part of code includes 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 from that marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved.

[0036] If the above functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical disks, and other various media that can store program codes.

[0037] The specific embodiments described above further elaborate on the purpose, technical solution, and beneficial effects of the present application. It should be understood that the above description is only for the specific embodiments of the present application and is not used to limit the protection scope of the present application. It is particularly pointed out that for those skilled in the art, any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for detecting multiple antibiotics in water based on artificial intelligence, characterized in that: include: The water body to be tested is processed by a solid phase extraction column, and the spectral data of the water body to be tested is detected by a high performance liquid chromatography tandem mass spectrometer; Based on the spectral data, extracting multidimensional feature data including retention time, parent ion mass-to-charge ratio, fragment ion mass-to-charge ratio, fragment ion intensity, peak symmetry parameters and signal-to-noise ratio; Based on the multidimensional feature data, construct a multidimensional feature matrix; A pre-trained deep learning model is used to perform multi-target recognition on the multidimensional feature matrix, and qualitative and quantitative analysis results of multiple antibiotics are output. The deep learning model is trained based on the spectral data of multiple antibiotic standard solution samples and the spectral data set of multiple water samples, and the spectral data set includes spectral data under different concentrations, different matrices and different interference conditions.

2. The method for detecting multiple antibiotics in water based on artificial intelligence as claimed in claim 1, characterized in that: The method of treating the water body to be tested by a solid phase extraction column and detecting the spectral data of the water body to be tested by a high performance liquid chromatography tandem mass spectrometer comprises: The water body to be tested was concentrated by adsorption on an HLB solid phase extraction column, eluted with methanol solution, blown to near dryness with nitrogen at 40°C, fixed to volume with 10% methanol aqueous solution, and detected in a high performance liquid chromatography tandem mass spectrometer using acetonitrile-5mmol / L ammonium acetate aqueous solution and acetonitrile-10mmol / L trifluoroacetic acid aqueous solution as mobile phases, respectively, to obtain the spectral data of the water body to be tested.

3. The method for detecting multiple antibiotics in water based on artificial intelligence as claimed in claim 1, characterized in that: Based on the spectral data, extracting multidimensional feature data including retention time, parent ion mass-to-charge ratio, fragment ion mass-to-charge ratio, fragment ion intensity, peak symmetry parameters and signal-to-noise ratio, including: Performing wavelet transformation and time alignment on the atlas data to obtain target atlas data; The target spectrum data is feature extracted using a preset spectrum feature extraction model to obtain multidimensional feature data including the retention time, parent ion mass-to-charge ratio, fragment ion mass-to-charge ratio, fragment ion intensity, peak symmetry parameters and signal-to-noise ratio. The preset spectrum feature extraction model is a model obtained by training a machine learning algorithm based on a preset spectrum training set.

4. The method for detecting multiple antibiotics in water based on artificial intelligence as claimed in claim 1, characterized in that: The step of constructing a multidimensional feature matrix based on the multidimensional feature data comprises: Calculate the center and interquartile range of each dimension of feature data in the multidimensional feature data; Using an anti-outlier standardization algorithm, according to the data center and the interquartile range, the original eigenvalue of each dimension of the eigendata is determined; Based on the original eigenvalues, median replacement or truncation processing is performed on the outliers in each dimension of the feature data to obtain target multidimensional feature data; The target multi-dimensional feature data are combined to construct the multi-dimensional feature matrix.

5. The method for detecting multiple antibiotics in water based on artificial intelligence as claimed in claim 1, characterized in that: The pre-trained deep learning model is used to perform multi-target recognition on the multi-dimensional feature matrix, and outputs qualitative analysis results and quantitative analysis results of multiple antibiotics, including: The multidimensional feature matrix is ​​input into the deep learning model, and the prediction results of multiple antibiotic types are output through the softmax function, and the concentration prediction results of multiple antibiotics are output through the linear regression layer. The deep learning model is a convolutional neural network, a recurrent neural network, a long short-term memory network, a graph neural network or a multi-task learning model.

6. The method for detecting multiple antibiotics in water based on artificial intelligence as claimed in claim 1, characterized in that: The antibiotics include sulfathiazole, sulfamethoxazole, sulfamethoxypyridazine, sulfisoxazole, trimethoprim, sulfacetamide, sulfapyridine, sulfadiazine, sulfamethoxazole, sulfamethoxazole, sulfamethoxazole, sulfamethoxazole, sulfamethoxazole, sulfamethoxazole, sulfamethoxazole, sulfamethoxazole, sulfamethoxazole, sulfamethoxazole, sulfamethoxazole, sulfamethoxazole, sulfamethoxazole, sulfamethoxazole, sulfamethoxazole, sulfamethoxazole, sulfamethoxazole, sulfachloropyridazine, sulfaquinoxaline, sulfamethoxazole ...

7. A system for detecting multiple antibiotics in water based on artificial intelligence, characterized in that: It comprises a solid phase extraction device, an ultra-high performance liquid chromatograph, a triple quadrupole mass spectrometer and a computer device, wherein the computer device is used to implement the method for detecting multiple antibiotics in water based on artificial intelligence as described in any one of claims 1 to 6.

Images