A method and system for detecting CT value in PCR
By using historical fluorescence datasets to train algorithm models, and automate CT value detection in PCR detection, the problems of inconsistency and error of CT value detection results in traditional methods are solved, and higher detection accuracy and reliability are achieved.
Patent Information
- Application Number
- CN202411108830.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-13
- Publication Date
- 2025-02-11
- Estimated Expiration
- 2044-08-13
AI Technical Summary
Traditional PCR detection methods rely on manual judgment or simple threshold settings, resulting in inconsistency and error in CT value detection results.
By collecting the historical fluorescence data set and its corresponding historical CT reference values, establishing an associated data set, and using the data set to train the algorithm model, perform data cleaning and prediction, we automate the detection of CT values.
It improves the accuracy and objectivity of the CT value detection results, reduces the subjectivity of human judgment, and enhances the reliability and accuracy of PCR detection.
Smart Images
Figure CN118747277B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of PCR detection, and in particular to a PCR CT value detection method and system. Background Art
[0002] Traditional PCR (polymerase chain reaction) detection methods may rely on manual judgment or simple threshold setting to determine the CT value (the CT value is the number of amplification cycles corresponding to when the fluorescence signal of the amplified product reaches the set fluorescence threshold). This method is easily affected by the operator's subjective judgment, resulting in inconsistency and errors in results under different laboratories or experimental conditions. Summary of the invention
[0003] The purpose of the present invention is to provide a PCR CT value detection method and system, and the technical problem to be solved is to improve the accuracy of the detection result.
[0004] The present invention is achieved through the following technical solutions:
[0005] The first aspect provides a PCR CT value detection method, comprising the following steps:
[0006] S100, acquiring a historical fluorescence data set and a historical CT reference value corresponding to each historical fluorescence data in the historical fluorescence data set; associating each historical fluorescence data in the historical fluorescence data set with the corresponding historical CT reference value to obtain an associated data set;
[0007] S200, performing data cleaning on the associated data set to obtain a feature data set;
[0008] S300, using the feature data set to train an algorithm model to obtain a trained algorithm model;
[0009] S400, after receiving the detection task, collecting the fluorescence data at the current moment, and performing data cleaning on the fluorescence data;
[0010] S500, inputting the cleaned fluorescence data at the current moment into the trained algorithm model to obtain a CT prediction value corresponding to the fluorescence data;
[0011] S600, calling a CT reference value and an absolute threshold, and comparing an absolute difference between the CT reference value and the CT prediction value with the absolute threshold;
[0012] S610: If the absolute difference between the CT reference value and the CT predicted value is greater than the absolute threshold, the fluorescence data is abnormal data;
[0013] S620: If the absolute difference between the CT reference value and the CT predicted value is less than or equal to the absolute threshold, the fluorescence data is normal data; and the CT predicted value is used as the output result of the algorithm model;
[0014] S621, calling the positive CT value interval and the negative CT value interval, and determining the detection result of the fluorescence data through the CT prediction value, the positive CT value interval and the negative CT value interval;
[0015] S622: If the CT prediction value is within the positive CT value interval, the detection result of the fluorescence data is positive;
[0016] S623: If the CT prediction value is within the negative CT value interval, the detection result of the fluorescence data is negative.
[0017] By collecting the above historical fluorescence data sets and their corresponding historical CT reference values and establishing associated data sets, the richness and accuracy of the training data are ensured. The full use of historical data helps the algorithm model learn more feature patterns and laws, thereby improving the accuracy of prediction. Before training the algorithm model, the historical associated data sets are cleaned to complete the historical associated data, remove noise, outliers and inconsistent data, improve the quality of the training data, and reduce interference factors in the algorithm model training process, thereby improving the prediction performance of the algorithm model. The historical fluorescence data is extracted from the cleaned feature data set, and the algorithm model is trained using the historical fluorescence data. By training the algorithm model, the relationship between the fluorescence data and the CT value is learned, thereby more accurately predicting the CT value of the fluorescence data. After receiving the detection task, the fluorescence data at the current moment is collected and cleaned, and then the cleaned data is input into the trained algorithm model for prediction to obtain the CT prediction value. The CT prediction value is combined with the specific judgment criteria to avoid the subjectivity and uncertainty of human judgment and improve the objectivity and accuracy of the detection results. Through the introduction and continuous optimization of the machine algorithm model, this method can more accurately predict the CT value of unknown samples, providing more reliable and accurate support for PCR detection.
[0018] Furthermore, in S300, the feature data set is divided into a training set and a test set;
[0019] S310, using the above training set to train the algorithm model to obtain a trained algorithm model;
[0020] S320: Use the above test set to test the trained algorithm model.
[0021] The algorithm model is trained using the training set to learn the mapping relationship between the fluorescence data in the training set and the CT reference value. The model parameters are optimized through continuous iteration to minimize the prediction error. During the training process, the algorithm model will gradually learn the complex patterns in the feature data, thereby improving its prediction ability. After the training is completed, the performance of the trained algorithm model is evaluated using the test set. The corresponding CT prediction value is obtained by inputting the fluorescence data in the test set into the trained algorithm model, and compared with the CT reference value in the test set to evaluate whether the algorithm model can accurately predict the CT value of the fluorescence data.
[0022] Furthermore, the specific steps of S310 include:
[0023] S311, inputting the historical fluorescence data in the training set into the algorithm model to obtain a CT prediction value corresponding to the historical fluorescence data;
[0024] S312, extracting historical CT reference values associated with the historical fluorescence data from the training set;
[0025] S313, calling an absolute threshold, and comparing an absolute difference between a historical CT reference value and a CT prediction value corresponding to the historical fluorescence data with the absolute threshold;
[0026] S314: If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is greater than the absolute threshold, the historical fluorescence data is abnormal data, the historical fluorescence data is removed, and the training set is updated;
[0027] S315. If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is less than or equal to the absolute threshold, the historical fluorescence data is normal data; and the CT prediction value corresponding to the historical fluorescence data is used as the output result of the algorithm model.
[0028] Each piece of historical fluorescence data in the above training set is used as input, calculated through the algorithm model, and the corresponding CT prediction value is output; in order to improve the prediction accuracy of the algorithm model, it is necessary to obtain the CT reference value corresponding to each piece of historical fluorescence data as a reference; an absolute threshold is set, and the absolute threshold is used to measure the acceptable error range between the CT prediction value and the CT reference value. By comparing the absolute difference between the CT reference value and the CT prediction value and the absolute threshold, the feature data with too large absolute difference is eliminated. The reason for the excessive absolute difference may be abnormal data caused by incorrect measurement, data entry error or other reasons. The abnormal feature data will interfere with the training effect of the algorithm model and reduce the accuracy of the prediction results.
[0029] Furthermore, the specific steps of S320 include:
[0030] S321, inputting the historical fluorescence data in the test set into the trained algorithm model to obtain a CT prediction value corresponding to the historical fluorescence data;
[0031] S322, extracting historical CT reference values associated with the historical fluorescence data from the test set;
[0032] S323, calling an absolute threshold, and comparing an absolute difference between a historical CT reference value and a CT prediction value corresponding to the historical fluorescence data with the absolute threshold;
[0033] S324: If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is greater than the absolute threshold, the historical fluorescence data is abnormal data, and the CT prediction value corresponding to the historical fluorescence data is marked as an abnormal value;
[0034] S325. If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is less than or equal to the absolute threshold, the historical fluorescence data is normal data; and the CT prediction value corresponding to the historical fluorescence data is used as the output result of the algorithm model.
[0035] The trained algorithm model is used to predict the historical fluorescence data in the test set to evaluate the performance of the algorithm model; each piece of historical fluorescence data in the test set is used as input, calculated through the trained algorithm model, and the corresponding CT prediction value is output; in order to verify the prediction accuracy of the algorithm model, the CT reference value corresponding to each piece of historical fluorescence data in the test set is obtained as a reference; the accuracy of the algorithm model prediction is evaluated by comparing the absolute difference between the CT prediction value and the CT reference value. Feature data with excessively large absolute differences are considered unreliable. The reason for the excessively large absolute difference may be that the algorithm model fails to fit well or there is a problem with the feature data itself; feature data with absolute differences within the absolute threshold range are considered reliable; the performance of the algorithm model can be evaluated based on the ratio of unreliable CT prediction values to reliable CT prediction values in the test set.
[0036] A second aspect provides a PCR CT value detection system, which uses the above detection method;
[0037] The detection system includes:
[0038] A collection module, the collection module is used to collect fluorescence data at the current moment;
[0039] A storage module, the storage module is connected to the acquisition module; the storage module is used to store historical fluorescence data sets, historical CT reference values corresponding to each historical fluorescence data in the historical fluorescence data sets, fluorescence data collected at the current moment, CT reference values, absolute thresholds, positive CT value intervals, and negative CT value intervals;
[0040] The data processing module is connected to the acquisition module and the storage module; the data processing module is used to perform the following operations:
[0041] S100, acquiring a historical fluorescence data set and a historical CT reference value corresponding to each historical fluorescence data in the historical fluorescence data set from the storage module; associating each historical fluorescence data in the historical fluorescence data set with the corresponding historical CT reference value to obtain an associated data set;
[0042] S200, performing data cleaning on the associated data set to obtain a feature data set;
[0043] S300, using the feature data set to train an algorithm model to obtain a trained algorithm model;
[0044] S400, after receiving the detection task, controlling the acquisition module to collect the fluorescence data at the current moment, and performing data cleaning on the fluorescence data;
[0045] S500, inputting the cleaned fluorescence data at the current moment into the trained algorithm model to obtain a CT prediction value corresponding to the fluorescence data;
[0046] S600, calling a CT reference value and an absolute threshold, and comparing an absolute difference between the CT reference value and the CT prediction value with the absolute threshold;
[0047] S610: If the absolute difference between the CT reference value and the CT predicted value is greater than the absolute threshold, the fluorescence data is abnormal data;
[0048] S620: If the absolute difference between the CT reference value and the CT predicted value is less than or equal to the absolute threshold, the fluorescence data is normal data; and the CT predicted value is used as the output result of the algorithm model;
[0049] S621, calling the positive CT value interval and the negative CT value interval, and determining the detection result of the fluorescence data through the CT prediction value, the positive CT value interval and the negative CT value interval;
[0050] S622: If the CT prediction value is within the positive CT value interval, the detection result of the fluorescence data is positive;
[0051] S623: If the CT prediction value is within the negative CT value interval, the detection result of the fluorescence data is negative.
[0052] Furthermore, the data processing module is also used to divide the feature data set into a training set and a test set in S300;
[0053] S310, using the above training set to train the algorithm model to obtain a trained algorithm model;
[0054] S320: Use the above test set to test the trained algorithm model.
[0055] Furthermore, the data processing module is further used to execute S310, and the specific steps of S310 include:
[0056] S311, inputting the historical fluorescence data in the training set into the algorithm model to obtain a CT prediction value corresponding to the historical fluorescence data;
[0057] S312, extracting historical CT reference values associated with the historical fluorescence data from the training set;
[0058] S313, calling an absolute threshold, and comparing an absolute difference between a historical CT reference value and a CT prediction value corresponding to the historical fluorescence data with the absolute threshold;
[0059] S314: If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is greater than the absolute threshold, the historical fluorescence data is abnormal data, the historical fluorescence data is removed, and the training set is updated;
[0060] S315. If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is less than or equal to the absolute threshold, the historical fluorescence data is normal data; and the CT prediction value corresponding to the historical fluorescence data is used as the output result of the algorithm model.
[0061] Furthermore, the data processing module is further used to execute S320, and the specific steps of S320 include:
[0062] S321, inputting the historical fluorescence data in the test set into the trained algorithm model to obtain a CT prediction value corresponding to the historical fluorescence data;
[0063] S322, extracting historical CT reference values associated with the historical fluorescence data from the test set;
[0064] S323, calling an absolute threshold, and comparing an absolute difference between a historical CT reference value and a CT prediction value corresponding to the historical fluorescence data with the absolute threshold;
[0065] S324: If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is greater than the absolute threshold, the historical fluorescence data is abnormal data, and the CT prediction value corresponding to the historical fluorescence data is marked as an abnormal value;
[0066] S325. If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is less than or equal to the absolute threshold, the historical fluorescence data is normal data; and the CT prediction value corresponding to the historical fluorescence data is used as the output result of the algorithm model.
[0067] Compared with the prior art, the present invention has the following advantages and beneficial effects:
[0068] By collecting the above historical fluorescence data sets and their corresponding historical CT reference values and establishing associated data sets, the richness and accuracy of the training data are ensured. The full use of historical data helps the algorithm model learn more feature patterns and laws, thereby improving the accuracy of prediction. Before training the algorithm model, the historical associated data sets are cleaned to complete the historical associated data, remove noise, outliers and inconsistent data, improve the quality of the training data, and reduce interference factors in the algorithm model training process, thereby improving the prediction performance of the algorithm model. The historical fluorescence data is extracted from the cleaned feature data set, and the algorithm model is trained using the historical fluorescence data. By training the algorithm model, the relationship between the fluorescence data and the CT value is learned, thereby more accurately predicting the CT value of the fluorescence data. After receiving the detection task, the fluorescence data at the current moment is collected and cleaned, and then the cleaned data is input into the trained algorithm model for prediction to obtain the CT prediction value. The CT prediction value is combined with the specific judgment criteria to avoid the subjectivity and uncertainty of human judgment and improve the objectivity and accuracy of the detection results. Through the introduction and continuous optimization of the machine algorithm model, this method can more accurately predict the CT value of unknown samples, providing more reliable and accurate support for PCR detection. BRIEF DESCRIPTION OF THE DRAWINGS
[0069] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:
[0070] Figure 1 As the main flow chart;
[0071] Figure 2 The CT prediction value, CT reference value and absolute difference value provided for Example 1;
[0072] Figure 3 The CT prediction value, CT reference value and absolute difference value provided for Example 2;
[0073] Figure 4CT predicted values, CT reference values, and absolute differences provided for Example 3. DETAILED DESCRIPTION
[0074] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.
[0075] Example 1
[0076] Combination Figure 1 This embodiment 1 provides a PCR CT value detection method, comprising the following steps:
[0077] S100, obtaining a historical fluorescence data set and a historical CT reference value corresponding to each historical fluorescence data in the historical fluorescence data set (the CT reference value is the CT value of the experimental instrument itself); associating each historical fluorescence data in the historical fluorescence data set with the corresponding historical CT reference value to obtain an associated data set;
[0078] S200, performing data cleaning on the associated data set to obtain a feature data set;
[0079] S300, using the feature data set to train an algorithm model (the algorithm model may be any one of XGBoost, RNN convolutional neural network, MLPRegressor, etc.) to obtain a trained algorithm model;
[0080] S400, after receiving the detection task, collecting the fluorescence data at the current moment, and performing data cleaning on the fluorescence data;
[0081] S500, inputting the cleaned fluorescence data at the current moment into the trained algorithm model to obtain a CT prediction value corresponding to the fluorescence data;
[0082] S600, calling the CT reference value and the absolute threshold value (the absolute threshold value can be adjusted according to the detected reagent), and comparing the absolute difference between the CT reference value and the CT predicted value with the absolute threshold value;
[0083] S610: If the absolute difference between the CT reference value and the CT predicted value is greater than the absolute threshold, the fluorescence data is abnormal data;
[0084] S620: If the absolute difference between the CT reference value and the CT predicted value is less than or equal to the absolute threshold, the fluorescence data is normal data; and the CT predicted value is used as the output result of the algorithm model;
[0085] S621, calling the positive CT value interval and the negative CT value interval, and determining the detection result of the fluorescence data through the CT prediction value, the positive CT value interval and the negative CT value interval;
[0086] S622: If the CT prediction value is within the positive CT value interval, the detection result of the fluorescence data is positive;
[0087] S623: If the CT prediction value is within the negative CT value interval, the detection result of the fluorescence data is negative.
[0088] By collecting the above-mentioned historical fluorescence data sets and their corresponding historical CT reference values and establishing associated data sets, the richness and accuracy of the training data are ensured. The full use of historical data helps the algorithm model learn more feature patterns and laws, thereby improving the accuracy of predictions. Before training the algorithm model, the historical associated data sets are cleaned to complete the historical associated data, remove noise, outliers and inconsistent data, improve the quality of the training data, and reduce interference factors in the algorithm model training process, thereby improving the prediction performance of the algorithm model. The historical fluorescence data is extracted from the cleaned feature data set, and the historical fluorescence data is used to train the algorithm model. By training the algorithm model, the learning The relationship between fluorescence data and CT value is learned, so as to more accurately predict the CT value of fluorescence data; after receiving the detection task, the fluorescence data at the current moment is collected and the data is cleaned, and then the cleaned data is input into the trained algorithm model for prediction to obtain the CT prediction value; the experimental abnormality is determined by the CT prediction value and the CT reference value, and the CT prediction value under normal circumstances is combined with the specific judgment standard, which avoids the subjectivity and uncertainty of human judgment and improves the objectivity and accuracy of the detection results; through the introduction and continuous optimization of the machine algorithm model, this method can more accurately predict the CT value of unknown samples, providing more reliable and accurate support for PCR detection.
[0089] Example 2
[0090] On the basis of Example 1, in S300, the feature data set is divided into a training set and a test set;
[0091] S310, using the above training set to train the algorithm model to obtain a trained algorithm model;
[0092] S320: Use the above test set to test the trained algorithm model.
[0093] The algorithm model is trained using the training set to learn the mapping relationship between the fluorescence data in the training set and the CT reference value. The model parameters are optimized through continuous iteration to minimize the prediction error. During the training process, the algorithm model will gradually learn the complex patterns in the feature data, thereby improving its prediction ability. After the training is completed, the performance of the trained algorithm model is evaluated using the test set. The corresponding CT prediction value is obtained by inputting the fluorescence data in the test set into the trained algorithm model, and compared with the CT reference value in the test set to evaluate whether the algorithm model can accurately predict the CT value of the fluorescence data.
[0094] Example 3
[0095] Based on Example 2, the specific steps of S310 include:
[0096] S311, inputting the historical fluorescence data in the training set into the algorithm model to obtain the CT prediction value corresponding to the historical fluorescence data;
[0097] S312, extracting historical CT reference values associated with the historical fluorescence data from the training set;
[0098] S313, calling an absolute threshold, and comparing an absolute difference between a historical CT reference value and a CT prediction value corresponding to the historical fluorescence data with the absolute threshold;
[0099] S314: If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is greater than the absolute threshold, the historical fluorescence data is abnormal data, and the historical fluorescence data is removed, and the training set is updated;
[0100] S315. If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is less than or equal to the absolute threshold, the historical fluorescence data is normal data; and the CT prediction value corresponding to the historical fluorescence data is used as the output result of the algorithm model.
[0101] Each piece of historical fluorescence data in the above training set is used as input, calculated through the algorithm model, and the corresponding CT prediction value is output; in order to improve the prediction accuracy of the algorithm model, it is necessary to obtain the CT reference value corresponding to each piece of historical fluorescence data as a reference; an absolute threshold is set, and the absolute threshold is used to measure the acceptable error range between the CT prediction value and the CT reference value. By comparing the absolute difference between the CT reference value and the CT prediction value and the absolute threshold, the feature data with too large absolute difference is eliminated. The reason for the excessive absolute difference may be abnormal data caused by incorrect measurement, data entry error or other reasons. The abnormal feature data will interfere with the training effect of the algorithm model and reduce the accuracy of the prediction results.
[0102] Example 4
[0103] Based on Example 2, the specific steps of S320 include:
[0104] S321, inputting the historical fluorescence data in the test set into the trained algorithm model to obtain a CT prediction value corresponding to the historical fluorescence data;
[0105] S322, extracting historical CT reference values associated with the historical fluorescence data from the test set;
[0106] S323, calling an absolute threshold, and comparing an absolute difference between a historical CT reference value and a CT prediction value corresponding to the historical fluorescence data with the absolute threshold;
[0107] S324: If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is greater than the absolute threshold, the historical fluorescence data is abnormal data, and the CT prediction value corresponding to the historical fluorescence data is marked as an abnormal value;
[0108] S325. If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is less than or equal to the absolute threshold, the historical fluorescence data is normal data; and the CT prediction value corresponding to the historical fluorescence data is used as the output result of the algorithm model.
[0109] The trained algorithm model is used to predict the historical fluorescence data in the test set to evaluate the performance of the algorithm model; each piece of historical fluorescence data in the test set is used as input, calculated through the trained algorithm model, and the corresponding CT prediction value is output; in order to verify the prediction accuracy of the algorithm model, the CT reference value corresponding to each piece of historical fluorescence data in the test set is obtained as a reference; the accuracy of the algorithm model prediction is evaluated by comparing the absolute difference between the CT prediction value and the CT reference value. Feature data with excessively large absolute differences are considered unreliable. The reason for the excessively large absolute difference may be that the algorithm model fails to fit well or there is a problem with the feature data itself; feature data with absolute differences within the absolute threshold range are considered reliable; the performance of the algorithm model can be evaluated based on the ratio of unreliable CT prediction values to reliable CT prediction values in the test set.
[0110] For example, the feature data set includes 11918 historical fluorescence data with positive results and 2032 historical fluorescence data with negative results; the following data divisions are used to train the algorithm model:
[0111] Example 1: 11918 historical fluorescence data with positive results are used as training sets, and 2032 historical fluorescence data with negative results are used as test sets. The prediction results are as follows: Figure 2 As shown;
[0112] Example 2: 9535 historical fluorescence data with positive results are used as training sets, 2032 historical fluorescence data with negative results and 2383 historical fluorescence data with positive results are used as test sets. The prediction results are as follows: Figure 3 As shown;
[0113] Example 3: 11160 historical fluorescence data in the entire feature data set are used as the training set, and 2790 historical fluorescence data in the entire feature data set are used as the test set. The prediction results are as follows: Figure 4 As shown;
[0114] In summary, the more complete the types of training sets are, the more accurate the prediction results are; the model can not only find the normal CT prediction value output by normal fluorescence data, but also find abnormal fluorescence data and mark the abnormal fluorescence data; it can also train the abnormal fluorescence data separately to identify the cause of the abnormal fluorescence data, making the test results more accurate and able to remind the testers of abnormal experiments.
[0115] Example 5
[0116] This embodiment 5 provides a PCR CT value detection system, which adopts the above detection method;
[0117] The detection system includes:
[0118] A collection module, the collection module is used to collect fluorescence data at the current moment;
[0119] A storage module, the storage module is connected to the acquisition module; the storage module is used to store historical fluorescence data sets, historical CT reference values corresponding to each historical fluorescence data in the historical fluorescence data sets, fluorescence data collected at the current moment, CT reference values, absolute thresholds, positive CT value intervals, and negative CT value intervals;
[0120] The data processing module is connected to the acquisition module and the storage module; the data processing module is used to perform the following operations:
[0121] S100, acquiring a historical fluorescence data set and a historical CT reference value corresponding to each historical fluorescence data in the historical fluorescence data set from the storage module; associating each historical fluorescence data in the historical fluorescence data set with the corresponding historical CT reference value to obtain an associated data set;
[0122] S200, performing data cleaning on the associated data set to obtain a feature data set;
[0123] S300, using the feature data set to train an algorithm model to obtain a trained algorithm model;
[0124] S400, after receiving the detection task, controlling the acquisition module to collect the fluorescence data at the current moment, and performing data cleaning on the fluorescence data;
[0125] S500, inputting the cleaned fluorescence data at the current moment into the trained algorithm model to obtain a CT prediction value corresponding to the fluorescence data;
[0126] S600, calling a CT reference value and an absolute threshold, and comparing an absolute difference between the CT reference value and the CT prediction value with the absolute threshold;
[0127] S610: If the absolute difference between the CT reference value and the CT predicted value is greater than the absolute threshold, the fluorescence data is abnormal data;
[0128] S620: If the absolute difference between the CT reference value and the CT predicted value is less than or equal to the absolute threshold, the fluorescence data is normal data; and the CT predicted value is used as the output result of the algorithm model;
[0129] S621, calling the positive CT value interval and the negative CT value interval, and determining the detection result of the fluorescence data through the CT prediction value, the positive CT value interval and the negative CT value interval;
[0130] S622: If the CT prediction value is within the positive CT value interval, the detection result of the fluorescence data is positive;
[0131] S623: If the CT prediction value is within the negative CT value interval, the detection result of the fluorescence data is negative.
[0132] In a specific embodiment, the data processing module is further used to divide the feature data set into a training set and a test set in S300;
[0133] S311, inputting the historical fluorescence data in the training set into the algorithm model to obtain a CT prediction value corresponding to the historical fluorescence data;
[0134] S312, extracting historical CT reference values associated with the historical fluorescence data from the training set;
[0135] S313, calling an absolute threshold, and comparing an absolute difference between a historical CT reference value and a CT prediction value corresponding to the historical fluorescence data with the absolute threshold;
[0136] S314: If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is greater than the absolute threshold, the historical fluorescence data is abnormal data, the historical fluorescence data is removed, and the training set is updated;
[0137] S315. If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is less than or equal to the absolute threshold, the historical fluorescence data is normal data; and the CT prediction value corresponding to the historical fluorescence data is used as the output result of the algorithm model.
[0138] In a specific embodiment, the data processing module is further used to execute S320. The specific steps of S320 include:
[0139] S321, inputting the historical fluorescence data in the test set into the trained algorithm model to obtain a CT prediction value corresponding to the historical fluorescence data;
[0140] S322, extracting historical CT reference values associated with the historical fluorescence data from the test set;
[0141] S323, calling an absolute threshold, and comparing an absolute difference between a historical CT reference value and a CT prediction value corresponding to the historical fluorescence data with the absolute threshold;
[0142] S324: If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is greater than the absolute threshold, the historical fluorescence data is abnormal data, and the CT prediction value corresponding to the historical fluorescence data is marked as an abnormal value;
[0143] S325. If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is less than or equal to the absolute threshold, the historical fluorescence data is normal data; and the CT prediction value corresponding to the historical fluorescence data is used as the output result of the algorithm model.
[0144] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A method for detecting the CT value of PCR, characterized in that: The following steps are involved: S100, acquiring a historical fluorescence data set and a historical CT reference value corresponding to each historical fluorescence data in the historical fluorescence data set; Associating each historical fluorescence data in the historical fluorescence data set with a corresponding historical CT reference value to obtain an associated data set; S200, performing data cleaning on the associated data set to obtain a feature data set; S300, using the feature data set to train an algorithm model to obtain a trained algorithm model; S400, after receiving the detection task, collecting the fluorescence data at the current moment, and performing data cleaning on the fluorescence data; S500, inputting the cleaned fluorescence data at the current moment into the trained algorithm model to obtain a CT prediction value corresponding to the fluorescence data; S600, calling a CT reference value and an absolute threshold, and comparing an absolute difference between the CT reference value and the CT prediction value with the absolute threshold; S610: If the absolute difference between the CT reference value and the CT predicted value is greater than the absolute threshold, the fluorescence data is abnormal data; analyzing the abnormal cause of the fluorescence data and indicating the abnormal situation; S620: If the absolute difference between the CT reference value and the CT predicted value is less than or equal to the absolute threshold, the fluorescence data is normal data; and the CT predicted value is used as the output result of the algorithm model; S621, calling the positive CT value interval and the negative CT value interval, and determining the detection result of the fluorescence data through the CT prediction value, the positive CT value interval and the negative CT value interval; S622: If the CT prediction value is within the positive CT value interval, the detection result of the fluorescence data is positive; S623: If the CT prediction value is within the negative CT value interval, the detection result of the fluorescence data is negative.
2. The CT value detection method of PCR according to claim 1, characterized in that: In S300, the feature data set is divided into a training set and a test set; S310, using the training set to train the algorithm model to obtain a trained algorithm model; S320: Use the test set to test the trained algorithm model.
3. The CT value detection method of PCR according to claim 2, characterized in that: The specific steps of S310 include: S311, inputting the historical fluorescence data in the training set into the algorithm model to obtain a CT prediction value corresponding to the historical fluorescence data; S312, extracting historical CT reference values associated with the historical fluorescence data from the training set; S313, calling an absolute threshold, and comparing an absolute difference between a historical CT reference value and a CT prediction value corresponding to the historical fluorescence data with the absolute threshold; S314: If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is greater than the absolute threshold, the historical fluorescence data is abnormal data, the historical fluorescence data is removed, and the training set is updated; S315. If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is less than or equal to the absolute threshold, the historical fluorescence data is normal data; and the CT prediction value corresponding to the historical fluorescence data is used as the output result of the algorithm model.
4. The method for detecting the CT value of PCR according to claim 2, characterized in that: The specific steps of S320 include: S321, inputting the historical fluorescence data in the test set into the trained algorithm model to obtain a CT prediction value corresponding to the historical fluorescence data; S322, extracting historical CT reference values associated with the historical fluorescence data from the test set; S323, calling an absolute threshold, and comparing an absolute difference between a historical CT reference value and a CT prediction value corresponding to the historical fluorescence data with the absolute threshold; S324: If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is greater than the absolute threshold, the historical fluorescence data is abnormal data, and the CT prediction value corresponding to the historical fluorescence data is marked as an abnormal value; S325. If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is less than or equal to the absolute threshold, the historical fluorescence data is normal data; and the CT prediction value corresponding to the historical fluorescence data is used as the output result of the algorithm model.
5. A PCR CT value detection system, characterized in that: The detection system adopts the detection method according to any one of claims 1 to 4; The detection system includes: A collection module, the collection module is used to collect fluorescence data at the current moment; A storage module, the storage module is connected to the acquisition module; the storage module is used to store the historical fluorescence data set, the historical CT reference value corresponding to each historical fluorescence data in the historical fluorescence data set, the fluorescence data collected at the current moment, the CT reference value, the absolute threshold, the positive CT value interval and the negative CT value interval; A data processing module, the data processing module is connected to the acquisition module and the storage module; the data processing module is used to perform the following operations: S100, acquiring a historical fluorescence data set and a historical CT reference value corresponding to each historical fluorescence data in the historical fluorescence data set from the storage module; associating each historical fluorescence data in the historical fluorescence data set with the corresponding historical CT reference value to obtain an associated data set; S200, performing data cleaning on the associated data set to obtain a feature data set; S300, using the feature data set to train an algorithm model to obtain a trained algorithm model; S400, after receiving the detection task, controlling the acquisition module to collect the fluorescence data at the current moment, and performing data cleaning on the fluorescence data; S500, inputting the cleaned fluorescence data at the current moment into the trained algorithm model to obtain a CT prediction value corresponding to the fluorescence data; S600, calling a CT reference value and an absolute threshold, and comparing an absolute difference between the CT reference value and the CT prediction value with the absolute threshold; S610: If the absolute difference between the CT reference value and the CT predicted value is greater than the absolute threshold, the fluorescence data is abnormal data; S620: If the absolute difference between the CT reference value and the CT predicted value is less than or equal to the absolute threshold, the fluorescence data is normal data; and the CT predicted value is used as the output result of the algorithm model; S621, calling the positive CT value interval and the negative CT value interval, and determining the detection result of the fluorescence data through the CT prediction value, the positive CT value interval and the negative CT value interval; S622: If the CT prediction value is within the positive CT value interval, the detection result of the fluorescence data is positive; S623: If the CT prediction value is within the negative CT value interval, the detection result of the fluorescence data is negative.
6. A PCR CT value detection system according to claim 5, characterized in that: The data processing module is also used to divide the feature data set into a training set and a test set in S300; S310, using the training set to train the algorithm model to obtain a trained algorithm model; S320: Use the test set to test the trained algorithm model.
7. A PCR CT value detection system according to claim 6, characterized in that: The data processing module is also used to execute S310, and the specific steps of S310 include: S311, inputting the historical fluorescence data in the training set into the algorithm model to obtain a CT prediction value corresponding to the historical fluorescence data; S312, extracting historical CT reference values associated with the historical fluorescence data from the training set; S313, calling an absolute threshold, and comparing an absolute difference between a historical CT reference value and a CT prediction value corresponding to the historical fluorescence data with the absolute threshold; S314: If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is greater than the absolute threshold, the historical fluorescence data is abnormal data, the historical fluorescence data is removed, and the training set is updated; S315. If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is less than or equal to the absolute threshold, the historical fluorescence data is normal data; and the CT prediction value corresponding to the historical fluorescence data is used as the output result of the algorithm model.
8. A PCR CT value detection system according to claim 6, characterized in that: The data processing module is also used to execute S320, and the specific steps of S320 include: S321, inputting the historical fluorescence data in the test set into the trained algorithm model to obtain a CT prediction value corresponding to the historical fluorescence data; S322, extracting historical CT reference values associated with the historical fluorescence data from the test set; S323, calling an absolute threshold, and comparing an absolute difference between a historical CT reference value and a CT prediction value corresponding to the historical fluorescence data with the absolute threshold; S324: If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is greater than the absolute threshold, the historical fluorescence data is abnormal data, and the CT prediction value corresponding to the historical fluorescence data is marked as an abnormal value; S325. If the absolute difference between the historical CT reference value and the CT prediction value corresponding to the historical fluorescence data is less than or equal to the absolute threshold, the historical fluorescence data is normal data; and the CT prediction value corresponding to the historical fluorescence data is used as the output result of the algorithm model.
Citation Information
Patent Citations
Data center energy efficiency anomaly detection method and system based on adaptive prediction interval and medium
CN116383773A