Valuable traditional Chinese medicinal material quality control and pesticide residue field inspection method, system and equipment based on width learning algorithm

Abstract

The invention discloses a valuable traditional Chinese medicinal material quality control and pesticide residue field inspection method, system and equipment based on a width learning algorithm, and belongs to the technical field of traditional Chinese medicinal material detection. Through the open ion source and the width learning algorithm, quality control and pesticide residue on-site inspection of the precious traditional Chinese medicinal materials are achieved, chemical reagent consumption and complex sample pretreatment steps are avoided, the time of the quality control and pesticide residue inspection method of the precious traditional Chinese medicinal materials is shortened, and the requirement for rapid on-site detection is met; the cost of quality control of precious traditional Chinese medicinal materials and on-site inspection of pesticide residues is reduced; through a wide learning algorithm, quality control and pesticide residue field inspection of precious traditional Chinese medicinal materials are realized, a large amount of data can be effectively processed, the accuracy and reliability of data analysis are improved, and the scientificity and operability of a detection result are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of Chinese herbal medicine detection, and particularly to a method, system and device for quality control of precious Chinese herbal medicines and on-site inspection of agricultural residues based on a width learning algorithm. Background Art

[0002] The existing quality control of precious Chinese herbal medicines and on-site inspection of agricultural residues mainly rely on chemical analysis methods in laboratories, such as high performance liquid chromatography (HPLC), gas chromatography (GC) and mass spectrometry (MS), etc., to provide accurate chemical composition analysis.

[0003] However, in the scenario of on-site detection, the existing methods cannot meet the requirements of rapid on-site detection, resulting in their inability to be applied to situations where urgent detection results are needed to ensure the quality and safety of Chinese herbal medicines. Moreover, the above methods require complex instrument operations and professional knowledge, which limits their popularization in non-professional environments. The high cost of purchasing and maintaining high-end instruments further increases the cost of quality control of precious Chinese herbal medicines and on-site inspection of agricultural residues. Summary of the Invention

[0004] To solve the problems of the existing technology, embodiments of the present invention provide a method, system and device for quality control of precious Chinese herbal medicines and on-site inspection of agricultural residues based on a width learning algorithm, including:

[0005] On the one hand, a method for quality control of precious Chinese herbal medicines and on-site inspection of agricultural residues based on a width learning algorithm is provided. The method includes:

[0006] Obtain multiple groups of Chinese herbal medicine samples to be tested, and for any one of the Chinese herbal medicine samples to be tested, perform the following steps;

[0007] Put the Chinese herbal medicine sample to be tested into the injection module;

[0008] In the injection module, ionize the Chinese herbal medicine sample to be tested through an open ion source;

[0009] Capture relevant data corresponding to the Chinese herbal medicine sample through a tandem mass spectrometry system;

[0010] Extract features from the relevant data to obtain feature data;

[0011] Analyze the feature data through a width learning algorithm to obtain a detection result corresponding to the Chinese herbal medicine sample to be tested;

[0012] After obtaining multiple detection results corresponding to all the Chinese herbal medicine samples to be tested, generate a quality control and agricultural residue detection result of the Chinese herbal medicine samples according to the multiple detection results.

[0013] Optionally, before putting the traditional Chinese medicine sample to be tested into the sampling module, the method further includes:

[0014] Performing pretreatment on the multiple groups of traditional Chinese medicine samples to be tested to ensure that each traditional Chinese medicine sample to be tested has the same weight and the same size;

[0015] Drying and grinding the multiple groups of traditional Chinese medicine samples to be tested.

[0016] Optionally, within the sampling module, ionizing the traditional Chinese medicine sample to be tested through an open ion source includes:

[0017] Setting the optimal voltage and optimal temperature of the open ion source;

[0018] Ionizing the traditional Chinese medicine sample to be tested through the open ion source at the optimal voltage and the optimal temperature.

[0019] Optionally, capturing relevant data corresponding to the traditional Chinese medicine sample through a tandem mass spectrometry system includes:

[0020] Setting the resolution and sensitivity corresponding to the traditional Chinese medicine;

[0021] The tandem mass spectrometry system captures the relevant data according to the resolution and the sensitivity.

[0022] Optionally, performing feature extraction on the relevant data to obtain feature data includes:

[0023] Removing outliers and background noise from the relevant data;

[0024] Performing normalization processing on the cleaned relevant data to obtain normalized relevant data;

[0025] Performing feature extraction on the normalized relevant data according to a feature extraction algorithm to obtain feature data.

[0026] Optionally, the method further includes:

[0027] Presetting an initial width learning algorithm;

[0028] Setting the learning rate and the number of iterations of the initial width learning algorithm;

[0029] Training the initial width learning algorithm through training text to obtain the width learning algorithm;

[0030] Performing cross-validation on the width learning algorithm.

[0031] Optionally, after training the initial width learning algorithm with the training text to obtain the width learning algorithm, the method further includes:

[0032] Optimizing the loss function of the width learning algorithm through regularization.

[0033] Optionally, the cross-validation of the width learning algorithm includes:

[0034] Obtaining the accuracy rate, recall rate, harmonic mean, precision, and ROC curve corresponding to the width learning algorithm through independent test data sets;

[0035] Performing cross-validation on the width learning algorithm through the accuracy rate, the recall rate, the harmonic mean, the precision, and the ROC curve.

[0036] On the other hand, a precious traditional Chinese medicine material quality control and pesticide residue on-site inspection system based on the width learning algorithm is provided. The system includes a sample disposal device and a detection device, wherein:

[0037] The sample disposal device is used to obtain multiple groups of traditional Chinese medicine materials to be tested. For any one of the traditional Chinese medicine materials to be tested, the following steps are performed;

[0038] The sample disposal device is also used to put the traditional Chinese medicine material to be tested into the sampling module;

[0039] The sample disposal device is also used to ionize the traditional Chinese medicine material to be tested through an open ion source in the sampling module;

[0040] The sample disposal device is also used to capture relevant data corresponding to the traditional Chinese medicine material sample through a tandem mass spectrometry system;

[0041] The detection device is used to extract features from the relevant data to obtain feature data;

[0042] The detection device is also used to analyze the feature data through the width learning algorithm to obtain a detection result corresponding to the traditional Chinese medicine material sample to be tested;

[0043] The detection device is also used to generate a traditional Chinese medicine material quality control and pesticide residue detection result of the traditional Chinese medicine material sample according to the multiple detection results after obtaining multiple detection results corresponding to all the traditional Chinese medicine materials to be tested.

[0044] On the other hand, a precious traditional Chinese medicine material quality control and pesticide residue on-site inspection device based on the width learning algorithm is provided. The device includes:

[0045] A feature extraction module for extracting features from the relevant data to obtain feature data; the relevant data is obtained through an open ion source and a tandem mass spectrometry system;

[0046] An analysis module for analyzing the feature data through a width learning algorithm to obtain a detection result corresponding to the traditional Chinese medicine sample to be tested;

[0047] A detection result generation module for generating a quality control and pesticide residue detection result of the traditional Chinese medicine sample after obtaining multiple detection results corresponding to all the traditional Chinese medicine samples to be tested according to the multiple detection results.

[0048] The present invention has at least the following beneficial effects:

[0049] 1. Through the open ion source and the width learning algorithm, the quality control of precious traditional Chinese medicine and on-site pesticide residue inspection are realized, avoiding the consumption of chemical reagents and complex sample pretreatment steps, shortening the time of the quality control and pesticide residue inspection methods for precious traditional Chinese medicine, meeting the requirements of rapid on-site detection, and reducing the cost of on-site quality control and pesticide residue inspection of precious traditional Chinese medicine;

[0050] 2. Through the width learning algorithm, the quality control of precious traditional Chinese medicine and on-site pesticide residue inspection are realized, which can effectively process a large amount of data, improve the accuracy and reliability of data analysis, and ensure the scientificity and operability of the detection results. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0052] Figure 1 It is a schematic flowchart of a method for on-site quality control and pesticide residue inspection of precious traditional Chinese medicine based on the width learning algorithm provided by an embodiment of the present invention;

[0053] Figure 2 It is a schematic diagram of a system for on-site quality control and pesticide residue inspection of precious traditional Chinese medicine based on the width learning algorithm provided by an embodiment of the present invention;

[0054] Figure 3 It is a schematic diagram of the structure of a device for on-site quality control and pesticide residue inspection of precious traditional Chinese medicine based on the width learning algorithm provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0055] To make the objectives, technical solutions, and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0056] Refer to Figure 1 As shown, a method for quality control of valuable traditional Chinese medicine materials and on-site inspection of agricultural residues based on a width learning algorithm is provided. The method includes:

[0057] 101. Obtain multiple groups of traditional Chinese medicine materials to be tested;

[0058] For any one of the traditional Chinese medicine materials to be tested, perform the following steps;

[0059] 102. Place the traditional Chinese medicine materials to be tested into the sample injection module;

[0060] 103. In the sample injection module, ionize the traditional Chinese medicine materials to be tested through an open ion source;

[0061] 104. Through a tandem mass spectrometry system, capture relevant data corresponding to the traditional Chinese medicine materials samples;

[0062] 105. Extract features from the relevant data to obtain feature data;

[0063] 106. Analyze the feature data through a width learning algorithm to obtain the detection results corresponding to the traditional Chinese medicine materials to be tested;

[0064] 107. After obtaining multiple detection results corresponding to all the traditional Chinese medicine materials to be tested, generate the quality control and agricultural residue detection results of the traditional Chinese medicine materials samples according to the multiple detection results.

[0065] It should be noted that in step 107, it is also necessary to set the visualization form of the quality control and agricultural residue detection results of the traditional Chinese medicine materials: select a suitable visualization tool and format to display the analysis results, such as charts, graphs, etc., and it is also necessary to set the format and detail level of the generation of the quality control and agricultural residue detection results of the traditional Chinese medicine materials: determine the detail level and the information content included in the quality control and agricultural residue detection results of the traditional Chinese medicine materials to meet the needs of different users.

[0066] Optionally, before placing the traditional Chinese medicine materials to be tested into the sample injection module, the method further includes:

[0067] Preprocess multiple groups of traditional Chinese medicine materials to be tested to ensure that the weight and size of each traditional Chinese medicine material to be tested are the same;

[0068] Dry and grind multiple groups of traditional Chinese medicine samples to be tested.

[0069] The above steps are the sample preparation stage. By ensuring the uniformity of the size and weight of the traditional Chinese medicine samples to be tested, the consistency and comparability of the data are guaranteed. Through pretreatment such as drying and grinding the traditional Chinese medicine samples to be tested and standardizing these treatment steps, the influence of operation variation on the final result is reduced.

[0070] By appropriately preparing the traditional Chinese medicine samples, operations including cleaning and adjusting to an appropriate size and weight are carried out to ensure the consistency and accuracy of subsequent tests.

[0071] Optionally, in the injection module, ionize the traditional Chinese medicine samples to be tested through an open ion source, including:

[0072] Set the optimal voltage and optimal temperature of the open ion source;

[0073] At the optimal voltage and optimal temperature, ionize the traditional Chinese medicine samples to be tested through the open ion source.

[0074] By controlling the voltage and temperature of the ion source to the optimized level, the ionization efficiency and the ion output of the sample are improved.

[0075] Optionally, capture relevant data corresponding to the traditional Chinese medicine samples through a tandem mass spectrometry system, including:

[0076] Set the resolution and sensitivity corresponding to the traditional Chinese medicine;

[0077] The tandem mass spectrometry system captures relevant data according to the resolution and sensitivity.

[0078] By setting appropriate resolution and sensitivity, sufficient molecular information is ensured to be captured. Different settings may be required for different compounds.

[0079] Optionally, perform feature extraction on the relevant data to obtain feature data, including:

[0080] Remove outliers and background noise from the relevant data;

[0081] Perform normalization processing on the cleaned relevant data to obtain the normalized relevant data;

[0082] By removing outliers and background noise, the accuracy and cleanliness of the data are ensured. Further, by performing steps such as normalization or standardization, data under different batches or different conditions are processed for comparison, thereby improving the accuracy of the results.

[0083] According to the feature extraction algorithm, perform feature extraction on the normalized relevant data to obtain feature data, that is, identify and quantify key compounds in the sample by selecting an effective feature extraction method. The embodiments of the present invention do not limit the specific feature extraction algorithm.

[0084] It should be noted that the above process of removing outliers and background noise from the relevant data; normalizing the cleaned relevant data to obtain the normalized relevant data, and performing feature extraction on the normalized relevant data according to the feature extraction algorithm to obtain feature data can significantly improve the quality and efficiency of data analysis through data cleaning, normalization, and feature extraction, remove noise and outliers from the data, and improve the accuracy and reliability of the final model.

[0085] Furthermore, in the above process, the quality of the data is ensured by removing noise and outliers, making the model training more accurate, thereby improving the performance of the entire system.

[0086] By normalization and feature extraction, the complexity of the data model can be simplified, the data processing speed can be accelerated, and the model training and testing processes can be made more efficient.

[0087] In addition, it can also help the model better generalize to new and unseen data, and improve the stability and reliability of the model in practical applications.

[0088] Optionally, the method further includes:

[0089] Preset the initial width learning algorithm (Broad Learning System, BLS);

[0090] Set the learning rate and the number of iterations of the initial width learning algorithm;

[0091] Specifically, by selecting the specific parameter settings of the width learning algorithm, such as the learning rate and the number of iterations, these parameters have a significant impact on the performance and training speed of the model.

[0092] Train the initial width learning algorithm through the training text to obtain the width learning algorithm;

[0093] Perform cross-validation on the width learning algorithm, and adopt techniques such as cross-validation to evaluate the accuracy and generalization ability of the model.

[0094] Optionally, after training the initial width learning algorithm through the training text to obtain the width learning algorithm, the method further includes:

[0095] Optimize the loss function of the width learning algorithm through regularization;

[0096] In the process of optimizing the width learning algorithm, that is, in the process of optimizing the loss function of the width learning algorithm, the adjustment of the width learning algorithm structure and parameters is based on the results of previous training. The key to this process lies in accurately identifying and improving those aspects that perform poorly on the training dataset. To enhance the generalization ability of the width learning algorithm and prevent overfitting, regularization techniques will be adopted. Regularization limits the complexity of the width learning algorithm parameters by adding additional terms (such as L1 or L2 penalty terms) to the loss function, thereby reducing the risk of the width learning algorithm overfitting to the training data.

[0097] In addition, techniques such as dropout and early stopping may also be adopted. Dropout reduces overfitting by randomly ignoring some nodes in the neural network, while early stopping involves stopping training when the performance on the validation dataset stops improving.

[0098] These strategies work together to optimize the width learning algorithm structure and achieve more efficient and robust performance.

[0099] In addition, it should be noted that the core task of the validation stage of the width learning algorithm is to evaluate the performance of the model on unseen data, which is achieved by using an independent test dataset.

[0100] Optionally, cross-validating the width learning algorithm includes:

[0101] Through mutually independent test datasets, obtain the accuracy, recall rate, harmonic mean, precision, and ROC curve corresponding to the width learning algorithm;

[0102] Cross-validate the width learning algorithm through the accuracy, recall rate, harmonic mean, precision, and ROC curve.

[0103] The performance evaluation of the width learning algorithm is not limited to the accuracy rate, that is, the proportion of correct predictions of the width learning algorithm, but also includes the recall rate, that is, the ability of the width learning algorithm to correctly identify positive samples, and the F1 score (the harmonic mean of the accuracy rate and the recall rate, used to balance the two).

[0104] In addition, other metrics may also be evaluated, such as precision (that is, the proportion of actually positive samples among the samples predicted as positive by the width learning algorithm) and the area under the ROC curve (AUC, a method for measuring the classification performance of the model). Through these comprehensive metrics, the performance of the width learning algorithm in processing different types of data can be comprehensively evaluated, ensuring that the width learning algorithm not only performs well on a specific dataset but also has good generalization ability.

[0105] In summary, the method described in the embodiments of the present invention has the following effects;

[0106] Adopting the technical solution of the present invention can bring the following beneficial effects:

[0107] Through the application of an open ion source and a width learning algorithm, the present invention can significantly shorten the time for the quality and safety detection of traditional Chinese medicinal materials. This rapid detection ability is particularly suitable for occasions where immediate results are required, such as traditional Chinese medicinal material trading markets and on-site quality control, ensuring the rapid assessment of the quality and safety of traditional Chinese medicinal materials before entering the market or being used.

[0108] By simplifying the operation process, non-professionals can also easily perform detection operations. This is particularly important in reducing operation errors and improving data accuracy, and helps to promote it to a wider range of application scenarios.

[0109] Compared with traditional traditional Chinese medicinal material detection technologies (such as HPLC, GC-MS, etc.), the present invention reduces the consumption of expensive chemical reagents and complex sample pretreatment steps, and at the same time, the maintenance and operation costs of the equipment are also relatively low. This makes the detection of traditional Chinese medicinal materials more economical and affordable.

[0110] The width learning algorithm can effectively process a large amount of data, improving the accuracy and reliability of data analysis. This efficient data processing ability ensures the scientificity and operability of the detection results, and supports making more accurate quality control decisions.

[0111] It has good portability and adaptability, and is suitable for use in various environments, including fields, warehouses, and medicinal material processing sites, etc. This flexibility greatly expands the scope of use of the equipment, making the quality and safety detection of traditional Chinese medicinal materials not restricted by location.

[0112] Refer to Figure 2 As shown, a quality control and pesticide residue on-site inspection system for precious traditional Chinese medicinal materials based on a width learning algorithm is provided. The system includes a sample disposal device and a detection device, wherein:

[0113] The sample disposal device is used to obtain multiple groups of traditional Chinese medicinal material samples to be tested, and for any one of the traditional Chinese medicinal material samples to be tested, the following steps are performed;

[0114] The sample disposal device is also used to put the traditional Chinese medicinal material sample to be tested into the injection module;

[0115] The sample disposal device is also used to ionize the traditional Chinese medicinal material sample to be tested through an open ion source in the injection module;

[0116] The sample disposal device is also used to capture relevant data corresponding to the traditional Chinese medicinal material sample through a tandem mass spectrometry system;

[0117] The detection device is used to extract features from the relevant data to obtain feature data;

[0118] The detection device is also used to analyze the feature data through the width learning algorithm to obtain the detection results corresponding to the Chinese medicinal material samples to be tested;

[0119] After obtaining multiple detection results corresponding to all the Chinese medicinal material samples to be tested, the detection device is also used to generate the quality control and pesticide residue detection results of the Chinese medicinal material samples according to the multiple detection results.

[0120] Optionally, the sample handling device is also used to:

[0121] Preprocess multiple groups of Chinese medicinal material samples to be tested to ensure that each Chinese medicinal material sample to be tested has the same weight and the same size;

[0122] Dry and grind multiple groups of Chinese medicinal material samples to be tested.

[0123] Optionally, the sample handling device is also used to:

[0124] Set the optimal voltage and optimal temperature of the open ion source;

[0125] Ionize the Chinese medicinal material samples to be tested through the open ion source at the optimal voltage and optimal temperature.

[0126] Optionally, the sample handling device is also used to:

[0127] Set the resolution and sensitivity corresponding to the Chinese medicinal materials;

[0128] The tandem mass spectrometry system captures relevant data according to the resolution and sensitivity.

[0129] Optionally, the detection device is also used to:

[0130] Remove outliers and background noise from the relevant data;

[0131] Normalize the cleaned relevant data to obtain the normalized relevant data;

[0132] Extract features from the normalized relevant data according to the feature extraction algorithm to obtain feature data.

[0133] Optionally, the detection device is also used to:

[0134] Preset the initial width learning algorithm;

[0135] Set the learning rate and the number of iterations of the initial width learning algorithm;

[0136] Train the initial width learning algorithm through the training text to obtain the width learning algorithm;

[0137] Perform cross-validation on the width learning algorithm.

[0138] Optionally, the detection device is further configured to:

[0139] Optimize the loss function of the width learning algorithm through regularization.

[0140] Optionally, the detection device is further configured to:

[0141] Obtain the accuracy rate, recall rate, harmonic mean, precision, and ROC curve corresponding to the width learning algorithm through an independent test data set;

[0142] Perform cross-validation on the width learning algorithm through the accuracy rate, recall rate, harmonic mean, precision, and ROC curve.

[0143] Referring to Figure 3 As shown, a quality control and pesticide residue on-site inspection device for precious traditional Chinese medicine materials based on the width learning algorithm is provided. The quality control and pesticide residue on-site inspection device for precious traditional Chinese medicine materials includes:

[0144] A feature extraction module, configured to extract features from relevant data to obtain feature data; the relevant data is obtained through an open ion source and a tandem mass spectrometry system;

[0145] An analysis module, configured to analyze the feature data through the width learning algorithm to obtain a detection result corresponding to the traditional Chinese medicine material sample to be tested;

[0146] A detection result generation module, configured to generate a quality control and pesticide residue detection result of the traditional Chinese medicine material sample according to multiple detection results after obtaining multiple detection results corresponding to all traditional Chinese medicine material samples to be tested.

[0147] Optionally, the feature extraction module is specifically configured to:

[0148] Remove outliers and background noise from the relevant data;

[0149] Perform normalization processing on the cleaned relevant data to obtain normalized relevant data;

[0150] Extract features from the normalized relevant data according to the feature extraction algorithm to obtain feature data.

[0151] Optionally, the analysis module is further configured to:

[0152] Preset an initial width learning algorithm;

[0153] Set the learning rate and the number of iterations of the initial width learning algorithm;

[0154] Train the initial width learning algorithm through training text to obtain a width learning algorithm;

[0155] Perform cross-validation on the width learning algorithm.

[0156] Optionally, the analysis module is further configured to:

[0157] Optimize the loss function of the width learning algorithm through regularization.

[0158] Optionally, the analysis module is further configured to:

[0159] Obtain the accuracy rate, recall rate, harmonic mean, precision, and ROC curve corresponding to the width learning algorithm through an independent test data set;

[0160] Perform cross-validation on the width learning algorithm through the accuracy rate, recall rate, harmonic mean, precision, and ROC curve.

[0161] The above several specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in some embodiments.

[0162] The technical features of the above embodiments can be combined arbitrarily (as long as there is no contradiction in the combination of these technical features). For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described; these embodiments not explicitly written out should also be considered to be within the scope described in this specification.

[0163] In the above text, the present invention has been described in a relatively specific and detailed manner through general descriptions and specific embodiments. It should be noted that, without departing from the concept of the present invention, it is obvious that several modifications and improvements can still be made to these specific embodiments, and these all belong to the protection scope of this application. Therefore, the protection scope of the patent of this application should be subject to the appended claims.

[0164] The above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for on-site inspection of quality control and pesticide residues of precious traditional Chinese medicine materials based on the width learning algorithm, characterized in that, The method includes: Obtain multiple groups of traditional Chinese medicine (TCM) samples to be tested. For any one of the TCM samples to be tested, perform the following steps; Put the TCM sample to be tested into the sampling module; In the sampling module, ionize the TCM sample to be tested through an open ion source; Capture relevant data corresponding to the TCM sample through a tandem mass spectrometry system; Extract features from the relevant data to obtain feature data; Analyze the feature data through a width learning algorithm to obtain a detection result corresponding to the TCM sample to be tested; After obtaining multiple detection results corresponding to all TCM samples to be tested, generate a quality control and pesticide residue detection result of the TCM sample according to the multiple detection results.

2. The method according to claim 1, characterized in that Before putting the TCM sample to be tested into the sampling module, the method further includes: Pretreat the multiple groups of TCM samples to be tested to ensure that each TCM sample to be tested has the same weight and the same size; Dry and grind the multiple groups of TCM samples to be tested.

3. The method according to claim 2, wherein The ionizing the TCM sample to be tested through an open ion source in the sampling module includes: Set the optimal voltage and optimal temperature of the open ion source; At the optimal voltage and the optimal temperature, ionize the TCM sample to be tested through the open ion source.

4. The method according to claim 3, wherein The capturing relevant data corresponding to the TCM sample through a tandem mass spectrometry system includes: Set the resolution and sensitivity corresponding to the TCM; The tandem mass spectrometry system captures the relevant data according to the resolution and the sensitivity.

5. The method according to claim 4, characterized in that, The extracting features from the relevant data to obtain feature data includes: Remove outliers and background noise from the relevant data; Perform normalization processing on the cleaned relevant data to obtain normalized relevant data; According to a feature extraction algorithm, extract features from the normalized relevant data to obtain feature data.

6. The method according to claim 5, wherein The method further includes: Preset an initial width learning algorithm; Set the learning rate and the number of iterations of the initial width learning algorithm; Train the initial width learning algorithm through training text to obtain the width learning algorithm; Perform cross-validation on the width learning algorithm.

7. The method according to claim 6, wherein After training the initial width learning algorithm through training text to obtain the width learning algorithm, the method further includes: Optimize the loss function of the width learning algorithm through regularization.

8. The method according to claim 7, wherein The performing cross-validation on the width learning algorithm includes: Obtain the accuracy rate, recall rate, harmonic mean, precision, and ROC curve corresponding to the width learning algorithm through an independent test data set; Perform cross-validation on the width learning algorithm through the accuracy rate, the recall rate, the harmonic mean, the precision, and the ROC curve.

9. A precious traditional Chinese medicine material quality control and pesticide residue on-site inspection system based on the width learning algorithm, characterized in that, The system includes a sample handling device and a detection device, where: The sample handling device is used to obtain multiple groups of TCM samples to be tested. For any one of the TCM samples to be tested, perform the following steps; The sample handling device is further used to put the TCM sample to be tested into the sampling module; The sample disposal device is also used to ionize the traditional Chinese medicine sample to be tested in the sample injection module through an open ion source; The sample disposal device is also used to capture relevant data corresponding to the traditional Chinese medicine sample through a tandem mass spectrometry system; The detection device is used to extract features from the relevant data to obtain feature data; The detection device is also used to analyze the feature data through a width learning algorithm to obtain a detection result corresponding to the traditional Chinese medicine sample to be tested; The detection device is also used to generate a quality control and pesticide residue detection result of the traditional Chinese medicine sample according to the multiple detection results after obtaining the multiple detection results corresponding to all the traditional Chinese medicine samples to be tested.

10. A precious traditional Chinese medicine material quality control and pesticide residue on-site inspection device based on the width learning algorithm, characterized in that, The device includes: A feature extraction module, which is used to extract features from the relevant data to obtain feature data; the relevant data is obtained through an open ion source and a tandem mass spectrometry system; An analysis module, which is used to analyze the feature data through a width learning algorithm to obtain a detection result corresponding to the traditional Chinese medicine sample to be tested; A detection result generation module, which is used to generate a quality control and pesticide residue detection result of the traditional Chinese medicine sample according to the multiple detection results after obtaining the multiple detection results corresponding to all the traditional Chinese medicine samples to be tested.

Images