A data analysis method and device based on electroactive microorganism detection
By preprocessing and feature extraction of electroactive microbial detection data, multiple models are constructed and optimized, the problem of difficult identification and classification of electroactive microbial types in traditional methods is solved, and efficient and accurate identification of electroactive microbial types and Detection data analysis.
Patent Information
- Application Number
- CN202410766467.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-14
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-06-14
AI Technical Summary
Traditional electroactive microbial detection data analysis methods are inefficient and subjective, and are difficult to process data with high dimensions, high noise and many missing values, resulting in increased analysis difficulty and the inability to effectively identify and classify electroactive microbial types.
By preprocessing the electroactive microbial detection data, the power output-related characteristics and electrochemical characteristics are extracted, and prediction models, comprehensive characteristic analysis models, partial least squares regression spectroscopy analysis models, correlation models of gene sequences and metabolic activity and microbial classifications are constructed. The model is trained and optimized to achieve automated analysis.
It improves the accuracy and adaptability of electroactive microbial type identification and classification, enhances the accuracy and efficiency of detection data analysis, and can better capture the internal laws and characteristics of the data.
Smart Images

Figure CN118503787B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of data analysis, and in particular to a method and device for analyzing data based on electroactive microorganism detection. Background Art
[0002] With the rapid development of environmental monitoring and bioenergy, electroactive microbial detection technology has been widely used as an important technical means. However, due to the complexity and diversity of detection data, traditional data analysis methods often fail to meet actual needs. On the one hand, traditional data analysis methods mainly rely on manual experience, and have problems such as low efficiency and strong subjectivity; on the other hand, due to the high dimensionality, high noise, and many missing values of detection data, the difficulty of data analysis is greatly increased.
[0003] Therefore, how to automatically extract key features from massive electroactive microbial detection data and achieve efficient and accurate electroactive microbial type identification and classification, so as to effectively analyze the electroactive microbial detection data, is a technical problem that needs to be solved urgently. Summary of the invention
[0004] In order to achieve efficient and accurate identification and classification of electroactive microorganism types and analysis of detection data, and to improve the accuracy and adaptability of electroactive microorganism detection data analysis, the present application provides a data analysis method and device based on electroactive microorganism detection.
[0005] The above-mentioned invention objective of the present application is achieved through the following technical solutions:
[0006] A method for analyzing data based on electroactive microorganism detection comprises the following steps:
[0007] Pre-processing the collected raw data based on electroactive microorganism detection and the raw data based on environment and electrochemical system detection to obtain electrochemical system data, effective current data and input data set;
[0008] Conduct correlation analysis on electrochemical system data and effective current data, extract power output related features, and screen electrochemical characteristic data from electrochemical system data;
[0009] Construct prediction models, comprehensive characteristic analysis models, partial least squares regression spectral analysis models, gene sequence and metabolic activity association models, and microbial classification models based on input data sets, power output related characteristics, and electrochemical characteristic data;
[0010] The prediction model, comprehensive characteristic analysis model, partial least squares regression spectral analysis model, gene sequence and metabolic activity association model, and microbial classification model were trained and optimized respectively;
[0011] When the detection data of the electroactive microorganisms inputted by the user is received, the detection data is preprocessed based on a preset data processing flow, and the preprocessed detection data is inputted into the corresponding model for analysis.
[0012] By adopting the above technical scheme, the collected raw data on electroactive microbial detection and the raw data based on environmental and electrochemical system detection are first pre-processed to eliminate noise, outliers and missing values, etc., and reduce the impact on data analysis. Then, the electrochemical system data and the effective current data are correlated and analyzed to extract features related to power output. At the same time, the electrochemical characteristic data reflecting the electrochemical activity of microorganisms and revealing the interaction between microorganisms and the environment are screened from the electrochemical system data as the basis for subsequent model construction. Based on the input data set, power output-related features and electrochemical characteristic data, multiple models are constructed for predicting microbial activity, analyzing the relationship between microorganisms and the environment, spectral data analysis, gene-metabolism association analysis, and classifying microorganisms. After the model is built, each model is trained and optimized so that the model can better fit the data to improve the accuracy and generalization ability of the model, so that each model can accurately capture the inherent laws and characteristics of the data. After the training and optimization of the model are completed, the electroactive microorganism detection data input by the user end is preprocessed according to the preset data processing process, and the preprocessed detection data is input into the corresponding model for analysis to complete the detection data analysis based on electroactive microorganisms. The present application realizes automatic extraction of key features related to power output, etc. by preprocessing and correlation analysis of the collected original data of electroactive microorganism detection and the original data of environmental and electrochemical system detection, and constructs multiple models related to electroactive microorganism detection through the processed data, so as to realize efficient and accurate identification and classification of electroactive microorganism types and analysis of detection data, which has the effect of improving the accuracy and adaptability of electroactive microorganism detection data analysis.
[0013] In a preferred example, the present application can be further configured as follows: the step of pre-processing the collected raw data based on electroactive microorganism detection and the raw data based on environment and electrochemical system detection to obtain electrochemical system data, effective current data and input data set includes the steps of:
[0014] Collect the raw current data obtained from the detection of electroactive microorganisms, obtain the mixed signal including the electrical signal generated by microbial metabolism and environmental noise, and collect the experimental data of the electrochemical system;
[0015] Performing noise reduction processing on the original current data to remove clutter and interference signals to obtain noise-reduced current data;
[0016] The noise-reduced current data and the experimental data of the electrochemical system are normalized to obtain the effective current data and the electrochemical system data.
[0017] By adopting the above technical scheme, the original current signal obtained from the detection of electroactive microorganisms and the experimental data of the electrochemical system are collected to obtain the mixed signal including the electrical signal generated by microbial metabolism and the environmental noise and the information of the electrochemical system. The original current data is denoised to remove clutter and interference signals. The denoised current data and the experimental data of the electrochemical system are normalized to obtain effective current data and electrochemical system data as the basis for subsequent data analysis and model construction.
[0018] In a preferred example, the present application can be further configured as follows: the steps of performing correlation analysis on the electrochemical system data and the effective current data, extracting power output related features, and screening electrochemical feature data from the electrochemical system data include the steps of:
[0019] Calculating the correlation coefficients between each parameter of the effective current data, other preset parameters in the electrochemical system data and the power output;
[0020] If the correlation coefficient between any parameter and the power output is higher than a predetermined threshold, the parameter is determined to be a power output related feature.
[0021] By adopting the above technical scheme, the correlation coefficients between the parameters of the effective current data and the power output and the other preset parameters in the electrochemical system data except the power output and the power output are calculated respectively. When the correlation coefficient between any parameter or any multiple parameters and the power output is greater than a predetermined threshold, the parameter is determined as a relevant feature that affects the power output, and the automatic extraction of key features is completed.
[0022] In a preferred example, the present application can be further configured as follows: the input data set includes biofilm activity data, microbial gene sequence data, metabolite data, water sample data, colony characteristic data, electron transfer information data and spectral characteristic data, and the steps of constructing a prediction model, a comprehensive characteristic analysis model, a partial least squares regression spectral analysis model, a gene sequence and metabolic activity association model and a microbial classification model based on the input data set, power output related characteristics and electrochemical characteristic data include the steps of:
[0023] Constructing a prediction model based on the relevant features to predict the power output under different conditions based on the relevant features;
[0024] Perform dimensionality reduction on water sample data and electrochemical characteristic data, and build a comprehensive characteristic analysis model;
[0025] A partial least squares regression spectral analysis model was constructed based on biofilm activity data and spectral characteristic data;
[0026] Construct a correlation model between gene sequence and metabolic activity based on microbial gene sequence data and metabolite data;
[0027] A microbial classification model is constructed based on electronic transfer information data and colony characteristic data.
[0028] By adopting the above technical scheme, a prediction model for predicting power output under different conditions based on relevant characteristics is constructed based on relevant characteristics, a comprehensive characteristic analysis model for analyzing the combined influence of water samples and electrochemical characteristics is constructed based on water sample data after dimensionality reduction processing and electrochemical characteristic data, a partial least squares regression spectral analysis model for analyzing the relationship between biofilm activity and spectral characteristics is constructed based on biofilm activity data and spectral characteristic data, a correlation model for analyzing the correlation between gene sequence and metabolic activity is constructed based on microbial gene sequence data and metabolite data, and a microbial classification model for identifying and classifying microorganisms is constructed based on electron transfer information data and colony characteristic data. By constructing a variety of models based on electroactive microbial detection, efficient and accurate electroactive microbial type recognition and classification and detection data analysis are achieved, which has the effect of improving the accuracy and adaptability of electroactive microbial detection data analysis.
[0029] In a preferred example, the present application can be further configured as follows: the water sample data includes chemical oxygen demand data, and the steps of performing dimensionality reduction processing on the water sample data and the electrochemical characteristic data and constructing a comprehensive characteristic analysis model include the steps of:
[0030] Performing data cleaning and preprocessing on chemical oxygen demand data and electrochemical characteristic data, wherein the data cleaning and preprocessing include removing outliers and standardizing;
[0031] Based on the principal component analysis technology, the chemical oxygen demand data and electrochemical characteristic data after data cleaning and preprocessing are subjected to dimensionality reduction processing, several principal component characteristics are obtained, and a principal component loading matrix showing the correlation between each principal component and the chemical oxygen demand data and the electrochemical characteristic data is constructed;
[0032] Calculate the score of each sample on each principal component based on the principal component loading matrix and the chemical oxygen demand data and electrochemical characteristic data before data cleaning and preprocessing, wherein the sample is the chemical oxygen demand data and the corresponding electrochemical characteristic data in a single water sample data;
[0033] A comprehensive characteristic analysis model was constructed based on the scores of each sample on each principal component and the chemical oxygen demand data.
[0034] By adopting the above technical scheme, the chemical oxygen demand data and electrochemical characteristic data are cleaned and preprocessed, including removing outliers and standardizing. The cleaned and preprocessed data are subjected to dimensionality reduction processing based on the principal component analysis technology, and several principal component characteristics are obtained and a principal component load matrix is constructed. The principal component load matrix represents the weight of each original feature on each principal component. The score of each sample on each principal component is calculated based on the principal component load matrix result, and a comprehensive characteristic analysis model for analyzing the comprehensive influence of water samples and electrochemical characteristics is constructed based on the calculated scores and the original chemical oxygen demand data.
[0035] In a preferred example, the present application can be further configured as follows: the step of constructing a correlation model between gene sequence and metabolic activity based on microbial gene sequence data and metabolite data comprises the following steps:
[0036] Obtain the gene sequences of known microorganisms, and construct an evolutionary tree based on the microbial gene sequence data and the gene sequences of known microorganisms;
[0037] Analyze microbial gene sequence data through genetic variation analysis tools to obtain genetic characteristic data, and measure the gene expression levels of microorganisms under different conditions through transcriptomics methods to obtain gene regulation data;
[0038] A correlation model between gene sequence and metabolic activity is constructed based on microbial gene sequence data, metabolite data, evolutionary tree, genetic characteristic data and gene regulation data.
[0039] By adopting the above technical scheme, the gene sequence of known microorganisms is obtained and an evolutionary tree is constructed. The evolutionary tree can show the evolutionary relationship between microorganisms. The microbial gene sequence data is deeply analyzed by genetic variation analysis tools to obtain the genetic characteristic data of the microorganisms, and the gene expression level of the microorganism under different environmental conditions is determined by transcriptomics methods to obtain gene regulation data. Based on the microbial gene sequence data, metabolite data, evolutionary tree, genetic characteristic data and gene regulation data, an association model of gene sequence and metabolic activity is constructed to analyze the association between gene sequence and metabolic activity.
[0040] In a preferred example, the present application can be further configured as follows: the step of constructing a microbial classification model based on the electronically transmitted information data and the colony characteristic data comprises the following steps:
[0041] Data cleaning was performed on the electron transfer information data and colony characteristic data to obtain a comprehensive data set of extracellular electron transfer and colony characteristics;
[0042] Based on feature engineering technology, classification features related to the classification of electroactive microorganisms are extracted from the comprehensive data set, and the extracted classification features are scored using a preset scoring strategy;
[0043] Specific features within a specific score range are selected from the classification features through a preset screening method and used as input to the random forest model to build a microbial classification model.
[0044] By adopting the above technical solution, the electron transfer information data and colony characteristic data are cleaned to obtain a comprehensive data set of extracellular electron transfer and colony characteristics. The feature engineering technology is used to extract classification features related to the classification of electroactive microorganisms from the comprehensive data set, and the extracted classification features are scored. The specific features that meet the specific score range are screened out through a preset screening method and used as the input of the random forest model, thereby constructing a microbial classification model for identifying and classifying microorganisms.
[0045] The second object of the invention is achieved by the following technical solutions:
[0046] A device for analyzing data based on electroactive microorganism detection, comprising:
[0047] A data processing module is used to pre-process the collected raw data based on electroactive microorganism detection and the raw data based on environment and electrochemical system detection to obtain electrochemical system data, effective current data and input data set;
[0048] A feature extraction module is used to perform correlation analysis on the electrochemical system data and the effective current data, extract power output related features, and filter electrochemical feature data from the electrochemical system data;
[0049] Model building module, used to build prediction models, comprehensive characteristic analysis models, partial least squares regression spectral analysis models, gene sequence and metabolic activity association models, and microbial classification models based on input data sets, power output related characteristics, and electrochemical characteristic data;
[0050] Model optimization module, used to train and optimize the prediction model, comprehensive characteristic analysis model, partial least squares regression spectral analysis model, gene sequence and metabolic activity association model, and microbial classification model respectively;
[0051] The data input module is used to pre-process the detection data based on a preset data processing flow when receiving the detection data of electroactive microorganisms input by the user end, and input the pre-processed detection data into the corresponding model for analysis.
[0052] By adopting the above technical scheme, the data processing module is used to pre-process the collected raw data based on electroactive microorganism detection and the raw data based on environment and electrochemical system detection to obtain electrochemical system data, effective current data and input data set; the feature extraction module is used to perform correlation analysis on the electrochemical system data and the effective current data, extract power output related features, and screen electrochemical feature data from the electrochemical system data; the model construction module is used to construct a prediction model, a comprehensive characteristic analysis model, a partial least squares regression spectral analysis model, a gene sequence and metabolic activity association model and a microbial classification model based on the input data set, power output related features and electrochemical feature data; the model optimization module is used to train and optimize the prediction model, the comprehensive characteristic analysis model, the partial least squares regression spectral analysis model, the gene sequence and metabolic activity association model and the microbial classification model respectively; the data input module is used to pre-process the detection data based on a preset data processing flow when receiving the detection data of electroactive microorganisms input by the user end, and input the pre-processed detection data into the corresponding model for analysis.
[0053] In summary, the present application includes at least one of the following beneficial technical effects:
[0054] 1. This application performs pre-processing and correlation analysis on the collected raw data of electroactive microorganism detection and the raw data of environmental and electrochemical system detection, so as to automatically extract key features related to power output, etc., and construct multiple models related to electroactive microorganism detection through the processed data, so as to realize efficient and accurate electroactive microorganism type identification and classification and analysis of detection data, which has the effect of improving the accuracy and adaptability of electroactive microorganism detection data analysis;
[0055] 2. After the model construction is completed, this application trains and optimizes each model separately so that the model can better fit the data, so as to improve the accuracy and generalization ability of the model, so that each model can accurately capture the inherent laws and characteristics of the data, and further improve the accuracy and adaptability of each model to the analysis of electroactive microbial detection data. BRIEF DESCRIPTION OF THE DRAWINGS
[0056] Figure 1 This is a flow chart of an embodiment of a method for analyzing data based on electroactive microorganism detection in the present application;
[0057] Figure 2 This is a flowchart of an implementation of step S10 in an embodiment of a method for analyzing data based on electroactive microorganism detection in the present application;
[0058] Figure 3This is a flowchart of an implementation of step S30 in an embodiment of a method for analyzing data based on electroactive microorganism detection in the present application;
[0059] Figure 4 This is a flowchart of an implementation of step S32 in an embodiment of a method for analyzing data based on electroactive microorganism detection in the present application;
[0060] Figure 5 This is a flowchart of an implementation of step S34 in an embodiment of a method for analyzing data based on electroactive microorganism detection in the present application;
[0061] Figure 6 This is a flowchart for implementing step S35 in an embodiment of the method for analyzing data based on electroactive microorganism detection in the present application. DETAILED DESCRIPTION
[0062] The following is combined with Figure 1-6 This application is described in further detail.
[0063] In one embodiment, if Figure 1 As shown, the present application discloses a method for analyzing data based on electroactive microorganism detection, which specifically includes the following steps:
[0064] S10: pre-processing the collected raw data based on electroactive microorganism detection and the raw data based on environment and electrochemical system detection to obtain electrochemical system data, effective current data and input data set;
[0065] In this embodiment, the raw data based on electroactive microorganism detection is a set of raw data that is directly measured or observed by using electroactive microorganisms as biosensors or detectors without any processing or analysis, and is used to evaluate the activity, metabolic state, genetic composition, metabolites and electrochemical process-related characteristics of electroactive microorganisms; the raw data based on environmental and electrochemical system detection is the raw data about environmental monitoring (such as water quality, air quality, spectral characteristics, etc.) or electrochemical parameters (such as current, voltage, resistance, etc.) collected by environmental detection stations or electrochemical detection instruments; pre-processing is a series of pre-processing operations performed on the raw data before data analysis; electrochemical system data is data about the electrochemical system obtained by electrochemical methods or technologies, which reflects the state, performance or reaction process of the electrochemical system; effective current data is current data in the electrochemical system that can truly reflect the electrochemical process or system state after appropriate processing; the input data set is a set of other data except electrochemical system data and effective current data after pre-processing the collected raw data based on electroactive microorganism detection and the raw data based on environmental and electrochemical system detection.
[0066] Specifically, the collected raw data on electroactive microorganism detection and the raw data based on environment and electrochemical system detection are pre-processed to eliminate noise, outliers and missing values, reduce the impact on data analysis, and obtain electrochemical system data, effective current data and input data sets.
[0067] S20: performing correlation analysis on the electrochemical system data and the effective current data, extracting power output related features, and screening electrochemical characteristic data from the electrochemical system data;
[0068] In this embodiment, correlation analysis is a statistical method for studying the strength and direction of the relationship between multiple variables; extraction of power output related features is to extract parameters or indicators closely related to power output from raw data; electrochemical characteristic data is data or parameters in electrochemical system data that are closely related to electrochemical processes or system performance.
[0069] Specifically, a correlation analysis is performed on the electrochemical system data and the effective current data to determine the correlation between the parameters of the effective current data and the remaining parameters in the electrochemical system data and the power output, and closely related parameters are determined as power output-related features for extraction, and parameters closely related to the electrochemical process are screened from the electrochemical system data and determined as electrochemical feature data.
[0070] S30: Construct prediction models, comprehensive characteristic analysis models, partial least squares regression spectral analysis models, gene sequence and metabolic activity association models, and microbial classification models based on input data sets, power output related characteristics, and electrochemical characteristic data;
[0071] In this embodiment, the prediction model is a mathematical model that predicts the target variable or result based on the input data set and related features, specifically predicting the power output under different conditions based on the relevant features; the comprehensive characteristic analysis model is a model that comprehensively considers multiple electrochemical characteristic data and related features to analyze the overall performance or behavior of the electrochemical system, specifically for analyzing the comprehensive impact of water samples and electrochemical characteristics; the partial least squares regression spectral analysis model is a model that combines spectral data and target variables to establish the relationship between spectral features and target variables to achieve quantitative or qualitative analysis of spectral data, specifically for analyzing the relationship between biofilm activity and spectral characteristics; the association model of gene sequence and metabolic activity is a model for establishing association rules between gene sequence features and metabolic activity, specifically for analyzing the association between gene sequence and metabolic activity; the microbial classification model is a model for classifying microorganisms based on their characteristic data.
[0072] Specifically, based on the input data set, power output related characteristics and electrochemical characteristic data, a prediction model for predicting the power output under different conditions based on the relevant characteristics, a comprehensive characteristic analysis model for analyzing the combined influence of water samples and electrochemical characteristics, a partial least squares regression spectral analysis model for analyzing the relationship between biofilm activity and spectral characteristics, a gene sequence and metabolic activity association model for analyzing the association between gene sequence and metabolic activity, and a microbial classification model for classifying microorganisms are constructed.
[0073] S40: training and optimizing the prediction model, comprehensive characteristic analysis model, partial least squares regression spectral analysis model, gene sequence and metabolic activity association model, and microbial classification model respectively;
[0074] In this embodiment, the training process of the prediction model includes using known input data and corresponding power output to train the prediction model, which is usually implemented by a machine learning algorithm, such as linear regression, decision tree, neural network, etc.;
[0075] Among them, the optimization of the prediction model includes adjusting the parameters or structure of the prediction model to minimize the prediction error or improve the processing performance of the model on unseen data, which usually involves hyperparameter tuning, feature selection, model selection and other technologies. The training of the comprehensive characteristic analysis model includes finding the most representative system performance indicators based on weight allocation, feature combination or selection; the optimization of the comprehensive characteristic analysis model includes adding new features, adjusting the weights between features or optimizing the mathematical expression of the model. The training of the partial least squares regression spectral analysis model includes training using spectral data and corresponding target variables (such as concentration, activity, etc.) to obtain the best linear relationship between spectral data and target variables; the optimization of the partial least squares regression spectral analysis model includes optimizing parameters, such as the number of principal components, regularization parameters, etc., to improve the prediction ability and generalization performance of the model. The training of the association model between gene sequence and metabolic activity includes training with the gene sequence data of microorganisms and the corresponding metabolic activity data to obtain key regions or patterns in the gene sequence that are significantly associated with metabolic activity; the optimization of the association model between gene sequence and metabolic activity includes optimizing the parameters and algorithms of the association model to improve the accuracy of the model in predicting the metabolic activity of new samples, including feature selection, algorithm adjustment or data enhancement technology, etc. The training of the microbial classification model includes training with microbial feature data with category labels (such as gene sequence, metabolite spectrum, electrochemical characteristics, etc.) to obtain features or patterns that can distinguish different microbial categories; the optimization of the microbial classification model includes optimizing the parameters, structure or algorithm of the classification model to improve the classification accuracy and generalization ability, including adjusting the hyperparameters of the classifier, using more complex feature extraction methods or trying different classification algorithms.
[0076] Furthermore, the optimization of the partial least squares regression spectral analysis model also includes the preprocessing of spectral data (such as smoothing, baseline correction, etc.);
[0077] Furthermore, the optimization of the microbial classification model also includes the use of ensemble learning techniques (such as random forests, gradient boosting machines, etc.) to improve classification performance.
[0078] Specifically, the prediction model, comprehensive characteristic analysis model, partial least squares regression spectral analysis model, gene sequence and metabolic activity association model, and microbial classification model are trained and optimized based on different methods to improve the accuracy and generalization of each model.
[0079] S50: When the detection data of the electroactive microorganisms inputted by the user is received, the detection data is preprocessed based on a preset data processing flow, and the preprocessed detection data is inputted into a corresponding model for analysis.
[0080] In this embodiment, the preset data processing flow is a preset data preprocessing flow based on type matching of input detection data.
[0081] Specifically, after the model is built, trained and optimized, when the detection data of electroactive microorganisms input by the user is received, the type of the detection data is identified, and based on the identification result, it is preprocessed according to a preset data processing flow. After preprocessing, the detection data is input into the corresponding model for analysis.
[0082] In one embodiment, if Figure 2 As shown, step S10 includes the steps of:
[0083] S11: collecting raw current data obtained from the detection of electroactive microorganisms, obtaining a mixed signal including electrical signals generated by microbial metabolism and environmental noise, and collecting experimental data of the electrochemical system;
[0084] S12: performing noise reduction processing on the original current data to remove clutter and interference signals to obtain noise-reduced current data;
[0085] S13: normalizing the noise-reduced current data and the experimental data of the electrochemical system to obtain effective current data and electrochemical system data.
[0086] In this embodiment, the original current data is a mixed signal including electrical signals generated by microbial metabolism, environmental noise and system noise; noise reduction processing is the process of removing unnecessary clutter, interference signals or noise from the signal to improve the signal-to-noise ratio of the signal, specifically including filtering, wavelet transform, Fourier transform, etc.; normalization processing is to scale the data according to certain preset rules so that it falls within a specific range for comparison and analysis.
[0087] Specifically, the original current signal obtained from the electroactive microorganism detection and the experimental data of the electrochemical system are collected to obtain the mixed signal including the electrical signal generated by microbial metabolism and the environmental noise and the information of the electrochemical system. The original current data is denoised to remove clutter and interference signals. The denoised current data and the experimental data of the electrochemical system are normalized to obtain effective current data and electrochemical system data as the basis for subsequent data analysis and model construction.
[0088] Furthermore, the wavelet transform technology is used to perform noise reduction processing, including the steps of:
[0089] Based on the wavelet transform technology, the mixed signal containing the electrical signal generated by microbial metabolism and the environmental noise is analyzed in time and frequency. According to the wavelet transform results, the signal is decomposed into multiple scales to identify the high-frequency noise and the low-frequency useful signal.
[0090] Remove high-frequency noise and retain low-frequency useful signals;
[0091] The low-frequency useful signal is further optimized based on the signal processing algorithm to ensure its authenticity.
[0092] In one embodiment, step S20 includes the steps of:
[0093] S21: calculating the correlation coefficients between each parameter of the effective current data, other preset parameters in the electrochemical system data and the power output respectively;
[0094] S22: If the correlation coefficient between any parameter and the power output is higher than a predetermined threshold, the parameter is determined to be a power output related feature.
[0095] In this embodiment, the correlation coefficient is a statistic used to measure the correlation between each parameter in the effective current data and the power output in the electrochemical system data, and between each parameter other than the power output in the electrochemical system data and the power output.
[0096] Specifically, the correlation coefficients between the parameters of the effective current data and the power output and the other preset parameters in the electrochemical system data except the power output and the power output are calculated respectively. When the correlation coefficient between any parameter or any multiple parameters and the power output is greater than a predetermined threshold, the parameter is determined as a relevant feature affecting the power output, and the automatic extraction of key features is completed.
[0097] In one embodiment, the input data set includes biofilm activity data, microbial gene sequence data, metabolite data, water sample data, colony characteristic data, electron transfer information data, and spectral characteristic data, such as Figure 3 As shown, step S30 includes the steps of:
[0098] S31: constructing a prediction model for predicting power output under different conditions based on the relevant features based on the relevant features;
[0099] S32: Perform dimensionality reduction processing on water sample data and electrochemical characteristic data, and construct a comprehensive characteristic analysis model;
[0100] S33: Construct a partial least squares regression spectral analysis model based on biofilm activity data and spectral characteristic data;
[0101] S34: Construct a correlation model between gene sequence and metabolic activity based on microbial gene sequence data and metabolite data;
[0102] S35: Construct a microbial classification model based on electronic transfer information data and colony characteristic data.
[0103] In this embodiment, dimensionality reduction processing is a process of converting original high-dimensional data into low-dimensional data through mathematical transformation, which specifically includes principal component analysis, linear discriminant analysis, t-SNE, etc.
[0104] Specifically, a prediction model for predicting power output under different conditions based on relevant characteristics is constructed based on relevant characteristics; a comprehensive characteristic analysis model for analyzing the combined influence of water samples and electrochemical characteristics is constructed based on water sample data after dimensionality reduction processing and electrochemical characteristic data; a partial least squares regression spectral analysis model for analyzing the relationship between biofilm activity and spectral characteristics is constructed based on biofilm activity data and spectral characteristic data; a correlation model for analyzing the correlation between gene sequence and metabolic activity is constructed based on microbial gene sequence data and metabolite data; a microbial classification model for identifying and classifying microorganisms is constructed based on electron transfer information data and colony characteristic data; and by constructing a variety of different models based on electroactive microbial detection, efficient and accurate identification and classification of electroactive microbial types and detection data analysis can be achieved.
[0105] In one embodiment, if Figure 4 As shown, the water sample data includes chemical oxygen demand data, and step S32 includes the steps of:
[0106] S321: performing data cleaning and preprocessing on the chemical oxygen demand data and the electrochemical characteristic data, wherein the data cleaning and preprocessing include removing abnormal values and standardizing;
[0107] S322: performing dimensionality reduction processing on the chemical oxygen demand data and the electrochemical characteristic data after data cleaning and preprocessing based on principal component analysis technology, obtaining a number of principal component characteristics and constructing a principal component load matrix showing the correlation between each principal component and the chemical oxygen demand data and the electrochemical characteristic data;
[0108] S323: Calculate the score of each sample on each principal component based on the principal component loading matrix and the chemical oxygen demand data and electrochemical characteristic data before data cleaning and preprocessing, wherein the sample is the chemical oxygen demand data and the corresponding electrochemical characteristic data in a single water sample data;
[0109] S324: Construct a comprehensive characteristic analysis model based on the score of each sample on each principal component and the chemical oxygen demand data.
[0110] In this embodiment, data cleaning and preprocessing are methods for removing interference data including removing outliers and standardization; principal component analysis is to transform the original data into a set of linearly unrelated variables (i.e., principal component features) through orthogonal transformation, and these principal component features can retain as much variation information in the original data as possible; the principal component loading matrix is an important output in principal component analysis, which represents the correlation between the original variables and the principal components.
[0111] Specifically, the chemical oxygen demand data and electrochemical characteristic data are cleaned and preprocessed, including removing outliers and standardizing. The cleaned and preprocessed data are subjected to dimensionality reduction based on principal component analysis technology to obtain several principal component features and construct a principal component loading matrix, which represents the weight of each original feature on each principal component. The score of each sample on each principal component is calculated based on the principal component loading matrix results, and a comprehensive characteristic analysis model for analyzing the comprehensive impact of water samples and electrochemical characteristics is constructed based on the calculated scores and the original chemical oxygen demand data.
[0112] Furthermore, the water sample data includes conductivity, pH value, redox potential, dissolved oxygen and chemical oxygen demand (COD); based on the principal component analysis technology, the conductivity, pH value, redox potential, dissolved oxygen and electrochemical characteristic data are analyzed to obtain several principal component characteristics.
[0113] Furthermore, the constructed comprehensive characteristic analysis model was subjected to goodness-of-fit analysis and significance test to determine the explanatory power and predictive accuracy of the model and to verify the effectiveness of the model.
[0114] In one embodiment, if Figure 5 As shown, step S34 includes the steps of:
[0115] S341: Obtain the gene sequence of a known microorganism, and construct an evolutionary tree based on the microbial gene sequence data and the gene sequence of the known microorganism;
[0116] S342: Analyze microbial gene sequence data using genetic variation analysis tools to obtain genetic characteristic data, and measure the gene expression levels of microorganisms under different conditions using transcriptomics methods to obtain gene regulation data;
[0117] S343: Construct a correlation model between gene sequence and metabolic activity based on microbial gene sequence data, metabolite data, evolutionary tree, genetic characteristic data, and gene regulation data.
[0118] In this embodiment, the gene sequence of a known microorganism is the nucleotide sequence of DNA or RNA in the genome of the microorganism, which contains the information required for life activities such as growth, reproduction, and metabolism of the microorganism; the evolutionary tree is a display of the evolutionary relationship of genes in the form of a tree diagram; the genetic variation analysis tool is a software or algorithm used to analyze genetic variation in the genome of an organism; transcriptomics is a disciplinary method for studying all RNA molecules of a specific cell or tissue under specific physiological or pathological conditions; and the gene regulation data is data describing the regulation mechanism of gene expression, including information on gene transcription levels, post-transcriptional regulation, and translation regulation.
[0119] Specifically, the gene sequences of known electroactive microorganisms are obtained and an evolutionary tree is constructed. The evolutionary tree can show the evolutionary relationship between electroactive microorganisms. The microbial gene sequence data are deeply analyzed through genetic variation analysis tools to obtain the genetic characteristic data of the microorganisms. The gene expression levels of the microorganisms under different environmental conditions are determined using transcriptomics methods to obtain gene regulation data. Based on the microbial gene sequence data, metabolite data, evolutionary tree, genetic characteristic data and gene regulation data, an association model of gene sequence and metabolic activity is constructed to analyze the association between gene sequence and metabolic activity.
[0120] In one embodiment, if Figure 6 As shown, step S35 includes the steps of:
[0121] S351: performing data cleaning on the electron transfer information data and the colony characteristic data to obtain a comprehensive data set of extracellular electron transfer and colony characteristics;
[0122] S352: Extract classification features related to the classification of electroactive microorganisms from the comprehensive data set based on feature engineering technology, and score the extracted classification features with a preset scoring strategy;
[0123] S353: Specific features within a specific score range are selected from the classification features through a preset screening method and used as input for a random forest model to construct a microbial classification model.
[0124] In this embodiment, the electron transfer information data is data describing the electron transfer activities of microorganisms (such as certain bacteria or fungi) outside the cell, wherein the electron transfer activities are usually related to the metabolic process, energy generation and transfer of the microorganisms; the colony characteristic data is data describing the size, shape, color, texture and other characteristics of the colonies formed by the growth of microorganisms on solid culture medium, which are often used for the classification and identification of microorganisms; data cleaning is a data preprocessing step to identify and correct errors, anomalies, missing values or inconsistent information in the data set to ensure the accuracy, completeness and consistency of the data; feature engineering technology is the process of extracting, constructing or selecting features from raw data, and the selected features help the machine learning model to better understand and predict the target variable; classification features are features closely related to classification tasks (such as microbial classification), and their features can help the machine learning model distinguish different categories; the preset scoring strategy is a method or standard for evaluating the importance of features in classification tasks. By scoring each feature, the impact of the feature on the performance of the classification model can be quantified; the random forest model is an ensemble learning method based on decision trees, which can improve the stability and accuracy of the model by constructing multiple decision trees and integrating their prediction results.
[0125] Specifically, the electron transfer information data and colony characteristic data are cleaned to obtain a comprehensive data set of extracellular electron transfer and colony characteristics. Feature engineering technology is used to extract classification features related to the classification of electroactive microorganisms from the comprehensive data set, and the extracted classification features are scored. Specific features that meet the specific score range are screened out through a preset screening method and used as input for the random forest model, thereby constructing a microbial classification model for identifying and classifying microorganisms.
[0126] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.
[0127] In one embodiment, a device for analyzing data based on electroactive microorganism detection is provided, and the device for analyzing data based on electroactive microorganism detection corresponds one-to-one to the method for analyzing data based on electroactive microorganism detection in the above embodiment.
[0128] A device for analyzing data based on electroactive microorganism detection, comprising:
[0129] A data processing module is used to pre-process the collected raw data based on electroactive microorganism detection and the raw data based on environment and electrochemical system detection to obtain electrochemical system data, effective current data and input data set;
[0130] A feature extraction module is used to perform correlation analysis on the electrochemical system data and the effective current data, extract power output related features, and filter electrochemical feature data from the electrochemical system data;
[0131] Model building module, used to build prediction models, comprehensive characteristic analysis models, partial least squares regression spectral analysis models, gene sequence and metabolic activity association models, and microbial classification models based on input data sets, power output related characteristics, and electrochemical characteristic data;
[0132] Model optimization module, used to train and optimize the prediction model, comprehensive characteristic analysis model, partial least squares regression spectral analysis model, gene sequence and metabolic activity association model, and microbial classification model respectively;
[0133] The data input module is used to pre-process the detection data based on a preset data processing flow when receiving the detection data of electroactive microorganisms input by the user end, and input the pre-processed detection data into the corresponding model for analysis.
[0134] Optionally, the data processing module includes:
[0135] The data acquisition submodule is used to collect the raw current data obtained from the detection of electroactive microorganisms, obtain the mixed signal including the electrical signal generated by microbial metabolism and the environmental noise, and collect the experimental data of the electrochemical system;
[0136] The noise reduction processing submodule is used to perform noise reduction processing on the original current data to remove clutter and interference signals and obtain noise-reduced current data;
[0137] The normalization processing submodule is used to normalize the noise-reduced current data and the experimental data of the electrochemical system to obtain effective current data and electrochemical system data.
[0138] Optionally, the feature extraction module includes:
[0139] The correlation coefficient calculation submodule is used to calculate the correlation coefficients between each parameter of the effective current data, other preset parameters in the electrochemical system data and the power output respectively;
[0140] The parameter comparison submodule is used to determine that any parameter is a power output related feature when the correlation coefficient between the parameter and the power output is higher than a predetermined threshold.
[0141] Optionally, model building modules include:
[0142] A prediction model building module, used for building a prediction model for predicting power output under different conditions based on relevant features;
[0143] A comprehensive characteristic analysis model building module is used to perform dimensionality reduction processing on water sample data and electrochemical characteristic data, and to build a comprehensive characteristic analysis model;
[0144] A partial least squares regression spectral analysis model building module is used to build a partial least squares regression spectral analysis model based on biofilm activity data and spectral characteristic data;
[0145] A module for building a model for the association between gene sequence and metabolic activity, which is used to build a model for the association between gene sequence and metabolic activity based on microbial gene sequence data and metabolite data;
[0146] The microbial classification model building module is used to build a microbial classification model based on electronic transmission information data and colony characteristic data.
[0147] Optional, comprehensive characterization model building blocks include:
[0148] A preprocessing submodule, used for cleaning and preprocessing the chemical oxygen demand data and the electrochemical characteristic data, wherein the data cleaning and preprocessing include removing abnormal values and standardizing;
[0149] The principal component load matrix construction submodule is used to perform dimensionality reduction processing on the chemical oxygen demand data and electrochemical characteristic data after data cleaning and preprocessing based on the principal component analysis technology, obtain several principal component characteristics and construct a principal component load matrix showing the correlation between each principal component and the chemical oxygen demand data and the electrochemical characteristic data;
[0150] A principal component score calculation submodule is used to calculate the score of each sample on each principal component based on the principal component loading matrix and the chemical oxygen demand data and electrochemical characteristic data before data cleaning and preprocessing;
[0151] The comprehensive characteristic analysis model construction submodule is used to construct a comprehensive characteristic analysis model based on the score of each sample on each principal component and the chemical oxygen demand data.
[0152] Optional modules for building a model for the association between gene sequence and metabolic activity include:
[0153] An evolutionary tree construction submodule is used to obtain the gene sequences of known microorganisms and construct an evolutionary tree based on the microbial gene sequence data and the gene sequences of known microorganisms;
[0154] The data acquisition submodule is used to analyze the microbial gene sequence data through genetic variation analysis tools to obtain genetic characteristic data, and to measure the gene expression levels of microorganisms under different conditions through transcriptomics methods to obtain gene regulation data;
[0155] The submodule for building a model for the association between gene sequence and metabolic activity is used to build a model for the association between gene sequence and metabolic activity based on microbial gene sequence data, metabolite data, evolutionary tree, genetic characteristic data, and gene regulation data.
[0156] Optional, microbial classification model building modules include:
[0157] The data cleaning submodule is used to clean the electron transfer information data and colony characteristic data to obtain a comprehensive data set of extracellular electron transfer and colony characteristics;
[0158] A feature scoring submodule is used to extract classification features related to the classification of electroactive microorganisms from the comprehensive data set based on feature engineering technology, and score the extracted classification features with a preset scoring strategy;
[0159] The microbial classification model construction submodule is used to select specific features within a specific score range from the classification features through a preset screening method, and use them as input for the random forest model to build a microbial classification model.
[0160] For the specific definition of a device for analyzing data based on electroactive microorganism detection, please refer to the definition of a method for analyzing data based on electroactive microorganism detection above, which will not be repeated here. Each module in the above-mentioned device for analyzing data based on electroactive microorganism detection can be implemented in whole or in part through software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.
[0161] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.
Claims
1. A method for analyzing data based on electroactive microorganism detection, characterized in that: Includes steps: Pre-processing the collected raw data based on electroactive microorganism detection and the raw data based on environment and electrochemical system detection to obtain electrochemical system data, effective current data and input data set; Conduct correlation analysis on electrochemical system data and effective current data, extract power output related features, and screen electrochemical characteristic data from electrochemical system data; Construct prediction models, comprehensive characteristic analysis models, partial least squares regression spectral analysis models, gene sequence and metabolic activity association models, and microbial classification models based on input data sets, power output related characteristics, and electrochemical characteristic data; The prediction model, comprehensive characteristic analysis model, partial least squares regression spectral analysis model, gene sequence and metabolic activity association model, and microbial classification model were trained and optimized respectively; When the detection data of the electroactive microorganisms inputted by the user is received, the detection data is preprocessed based on the preset data processing flow, and the preprocessed detection data is inputted into the corresponding model for analysis; The step of performing correlation analysis on the electrochemical system data and the effective current data, extracting power output related features, and screening electrochemical characteristic data from the electrochemical system data comprises the following steps: Calculating the correlation coefficients between each parameter of the effective current data, other preset parameters in the electrochemical system data and the power output; If the correlation coefficient between any parameter and the power output is higher than a predetermined threshold, the parameter is determined to be a power output related feature; The input data set includes biofilm activity data, microbial gene sequence data, metabolite data, water sample data, colony characteristic data, electron transfer information data and spectral characteristic data. The steps of constructing a prediction model, a comprehensive characteristic analysis model, a partial least squares regression spectral analysis model, a gene sequence and metabolic activity association model and a microbial classification model based on the input data set, power output related characteristics and electrochemical characteristic data include the following steps: Constructing a prediction model based on the relevant features to predict the power output under different conditions based on the relevant features; Perform dimensionality reduction on water sample data and electrochemical characteristic data, and build a comprehensive characteristic analysis model; A partial least squares regression spectral analysis model was constructed based on biofilm activity data and spectral characteristic data; Construct a correlation model between gene sequence and metabolic activity based on microbial gene sequence data and metabolite data; A microbial classification model is constructed based on electronic transfer information data and colony characteristic data.
2. The method for analyzing data based on electroactive microorganism detection according to claim 1, characterized in that: The step of pre-processing the collected raw data based on electroactive microorganism detection and the raw data based on environment and electrochemical system detection to obtain electrochemical system data, effective current data and input data set includes the steps of: Collect the raw current data obtained from the detection of electroactive microorganisms, obtain the mixed signal including the electrical signal generated by microbial metabolism and environmental noise, and collect the experimental data of the electrochemical system; Performing noise reduction processing on the original current data to remove clutter and interference signals to obtain noise-reduced current data; The noise-reduced current data and the experimental data of the electrochemical system are normalized to obtain the effective current data and the electrochemical system data.
3. The method for analyzing data based on electroactive microorganism detection according to claim 1, characterized in that: The water sample data includes chemical oxygen demand data. The steps of performing dimensionality reduction processing on the water sample data and the electrochemical characteristic data and constructing a comprehensive characteristic analysis model include the following steps: Performing data cleaning and preprocessing on chemical oxygen demand data and electrochemical characteristic data, wherein the data cleaning and preprocessing include removing outliers and standardizing; Based on the principal component analysis technology, the chemical oxygen demand data and electrochemical characteristic data after data cleaning and preprocessing are subjected to dimensionality reduction processing, several principal component characteristics are obtained, and a principal component loading matrix showing the correlation between each principal component and the chemical oxygen demand data and the electrochemical characteristic data is constructed; Calculate the score of each sample on each principal component based on the principal component loading matrix and the chemical oxygen demand data and electrochemical characteristic data before data cleaning and preprocessing, wherein the sample is the chemical oxygen demand data and the corresponding electrochemical characteristic data in a single water sample data; A comprehensive characteristic analysis model was constructed based on the scores of each sample on each principal component and the chemical oxygen demand data.
4. The method for analyzing data based on electroactive microorganism detection according to claim 1, characterized in that: The step of constructing a correlation model between gene sequence and metabolic activity based on microbial gene sequence data and metabolite data comprises the following steps: Obtain the gene sequences of known microorganisms, and construct an evolutionary tree based on the microbial gene sequence data and the gene sequences of known microorganisms; Analyze microbial gene sequence data through genetic variation analysis tools to obtain genetic characteristic data, and measure the gene expression levels of microorganisms under different conditions through transcriptomics methods to obtain gene regulation data; A correlation model between gene sequence and metabolic activity is constructed based on microbial gene sequence data, metabolite data, evolutionary tree, genetic characteristic data and gene regulation data.
5. The method for analyzing data based on electroactive microorganism detection according to claim 1, characterized in that: The step of constructing a microbial classification model based on electronically transmitted information data and colony characteristic data comprises the following steps: Data cleaning was performed on the electron transfer information data and colony characteristic data to obtain a comprehensive data set of extracellular electron transfer and colony characteristics; Based on feature engineering technology, classification features related to the classification of electroactive microorganisms are extracted from the comprehensive data set, and the extracted classification features are scored using a preset scoring strategy; Specific features within a specific score range are selected from the classification features through a preset screening method and used as input to the random forest model to build a microbial classification model.
6. A data analysis device based on electroactive microorganism detection, characterized in that: include: A data processing module is used to pre-process the collected raw data based on electroactive microorganism detection and the raw data based on environment and electrochemical system detection to obtain electrochemical system data, effective current data and input data set; A feature extraction module is used to perform correlation analysis on the electrochemical system data and the effective current data, extract power output related features, and filter electrochemical feature data from the electrochemical system data; Model building module, used to build prediction models, comprehensive characteristic analysis models, partial least squares regression spectral analysis models, gene sequence and metabolic activity association models, and microbial classification models based on input data sets, power output related characteristics, and electrochemical characteristic data; Model optimization module, used to train and optimize the prediction model, comprehensive characteristic analysis model, partial least squares regression spectral analysis model, gene sequence and metabolic activity association model, and microbial classification model respectively; A data input module is used to pre-process the detection data based on a preset data processing flow when receiving the detection data of electroactive microorganisms input by the user end, and input the pre-processed detection data into a corresponding model for analysis; The feature extraction module includes: The correlation coefficient calculation submodule is used to calculate the correlation coefficients between each parameter of the effective current data, other preset parameters in the electrochemical system data and the power output respectively; a parameter comparison submodule, for determining that any parameter is a power output-related feature when a correlation coefficient between the parameter and the power output is higher than a predetermined threshold; The input data set includes biofilm activity data, microbial gene sequence data, metabolite data, water sample data, colony characteristic data, electron transfer information data and spectral characteristic data. The model building modules include: A prediction model building module, used for building a prediction model for predicting power output under different conditions based on relevant features; A comprehensive characteristic analysis model building module is used to perform dimensionality reduction processing on water sample data and electrochemical characteristic data, and to build a comprehensive characteristic analysis model; A partial least squares regression spectral analysis model building module is used to build a partial least squares regression spectral analysis model based on biofilm activity data and spectral characteristic data; A module for building a model for the association between gene sequence and metabolic activity, which is used to build a model for the association between gene sequence and metabolic activity based on microbial gene sequence data and metabolite data; The microbial classification model building module is used to build a microbial classification model based on electronic transmission information data and colony characteristic data.
Citation Information
Patent Citations
Comprehensive poison monitoring method of microbial electrochemical sensor based on machine learning model
CN113820376A
Microorganism electrochemical analysis device and analysis method thereof
CN114813873A
Intelligent data early warning decision tree analysis method and system
CN117349782A
Ocean data processing and information scheduling system based on AI and Internet of Things technologies
CN117951584A
Microorganism culture result automatic analysis system
CN118051859A
