Microorganism data visualization chart generation system and method based on model parameters

Abstract

The invention relates to the technical field of visual chart generation, in particular to a microbial data visual chart generation system and method based on model parameters. According to the method, PCA dimensionality reduction and weighted LSTM modeling are combined, so that the processing efficiency and prediction precision of the microbial culture experimental data are improved. Through PCA dimension reduction, key features of a culture environment, microbial characteristics and a culture process are extracted, redundant information is reduced, and the model has generalization ability; on the basis of comparison between prediction of a weighted LSTM model and actual statistics of microbial biomass, an experiment trend is visualized, and optimization of culture conditions is facilitated; a real-time error calculation and anomaly detection mechanism can dynamically monitor the growth condition of microorganisms, accurately recognize anomaly time points, deeply analyze anomaly experiment parameters through an LOF algorithm, find out key anomaly components, finally generate a visual image and visually display anomaly sources and influence factors.

Description

Technical Field

[0001] The present invention relates to the technical field of visual chart generation, and specifically to a system and method for generating a visual chart of microbial data based on model parameters. Background Art

[0002] During the microbial culture process, minor changes in factors such as the culture environment, microbial characteristics, and operating conditions may lead to abnormal culture results, thereby affecting the accuracy of experiments and the stability of production. Therefore, it is crucial to detect and analyze abnormal data in the culture process in real time for optimizing culture conditions, improving the repeatability of experiments, and ensuring production quality. Traditional anomaly detection methods often rely on manual detection or simple threshold judgment, unable to track the dynamic growth of microorganisms in real time, easily resulting in anomalies not being discovered in time, and thus affecting experimental and production decisions.

[0003] Currently, the analysis of microbial culture data mainly uses static statistical methods such as mean analysis, variance analysis, and traditional regression models. Although these methods can provide certain reference information, they have many limitations. First, static analysis methods are usually based on data at fixed time points and are difficult to effectively reflect the complex dynamic changes in the microbial culture process. Second, existing analysis methods lack an intuitive data visualization presentation, making it difficult for researchers to quickly understand data trends and anomaly points. In addition, a single statistical analysis method is difficult to fully explore the potential information in experimental data, limiting the optimization space of the culture process. Therefore, there is an urgent need for a system that can detect anomalies in real time, perform dynamic modeling analysis, and intuitively display the changes in culture data through visualization means to improve the intelligent level of the microbial culture process.

[0004] Therefore, a system and method for generating a visual chart of microbial data based on model parameters are proposed. Summary of the Invention

[0005] The purpose of the present invention is to provide a system and method for generating a visual chart of microbial data based on model parameters. By using PCA dimensionality reduction, weighted LSTM modeling, and the Local Outlier Factor (LOF) algorithm, the quality, prediction accuracy, and anomaly detection ability of microbial culture data are improved. First, PCA dimensionality reduction optimizes historical experimental data, removes redundant information, retains key features, and improves calculation efficiency and data consistency. Second, based on the weighted LSTM model, accurate modeling is performed on the dimensionality-reduced experimental data. Different LSTM sub-modules are used to extract the core features of the culture environment, microbial characteristics, and culture process, and information fusion is optimized through dynamic weight allocation to achieve high-precision prediction of the number of microorganisms. Finally, the LOF algorithm combines KNN density estimation to analyze abnormal experimental parameters, accurately identify key abnormal components, and visually display the abnormal factors, providing a scientific basis for experimental optimization, anomaly traceability, and data correction.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A microbial data visualization chart generation system based on model parameters, comprising:

[0008] A principal component parameter acquisition module, configured to acquire historical experimental data of microbial culture, and perform dimensionality reduction on the historical experimental data by using the PCA method to obtain dimensionality-reduced experimental data; the historical experimental data includes historical experimental parameters and statistical microbial quantities;

[0009] A comparison data visualization module, configured to model the dimensionality-reduced experimental data by using a weighted LSTM model to obtain a pre-trained weighted LSTM model, record the test microbial quantity generated by a test set, and compare it with the statistical microbial quantity to generate a comparison visualization image;

[0010] An error visualization module, configured to record the initial microbial quantity, perform microbial quantity prediction and microbial quantity collection at every first time interval to obtain a predicted microbial quantity and a real-time microbial quantity, calculate the real-time error rate between the two, and generate an error visualization image;

[0011] An anomaly detection module, configured to judge the relationship between the real-time error rate and an error threshold. If the real-time error rate is greater than the error threshold, mark the calculation time of the real-time error rate as an anomaly point, and mark the anomaly point in the error visualization image;

[0012] An abnormal data visualization module, configured to obtain abnormal experimental parameters at an anomaly point, calculate the LOF value of each component of the abnormal experimental parameters by using the local outlier factor algorithm, obtain the key abnormal components of the abnormal experimental parameters, and output a visualization image of the key abnormal components.

[0013] Further, performing dimensionality reduction on the historical experimental data by using the PCA method includes:

[0014] Performing missing value filling, outlier removal, and standardization processing on the historical experimental parameters of the historical experimental data to obtain standardized experimental parameters;

[0015] Calculating the covariance matrix of the standardized experimental parameters, solving the eigenvalues and corresponding eigenvectors of the covariance matrix, and obtaining the principal component change direction;

[0016] Selecting the number of principal component spaces according to the cumulative variance contribution rate, and projecting the standardized experimental parameters into the selected principal component space to obtain principal component parameters.

[0017] Furthermore, the dimension-reduced experimental data includes principal component parameters and the statistical number of microorganisms, and the principal component parameters include culture environment parameters, microorganism characteristic parameters, and culture process parameters.

[0018] Furthermore, the weighted LSTM model includes:

[0019] A data preprocessing module for cleaning and normalizing the input data;

[0020] An LSTM module containing three LSTM sub-modules, which are respectively used to process the culture environment parameters, microorganism characteristic parameters, and culture process parameters;

[0021] A weight assignment module for assigning weights to the outputs of each LSTM sub-module;

[0022] A weighted fusion module: summing the outputs of different LSTM sub-modules with weights to obtain a weighted hidden state;

[0023] An output module for predicting the weighted hidden state and outputting the predicted microorganism quantity;

[0024] The calculation formula for the weighted hidden state is:

[0025] h final =ω env *h env +ω micro *h micro +ω process *h process ;

[0026] wherein, h final represents the weighted hidden state, h env represents the environmental hidden state, h micro represents the characteristic hidden state, h process represents the process hidden state, ω env represents the environmental weight coefficient, ω micro represents the characteristic weight coefficient, ω process represents the process weight coefficient.

[0027] Furthermore, the abnormal components for obtaining abnormal experimental parameters include:

[0028] Obtain abnormal experimental parameters, and combine the historical experimental parameters and the abnormal experimental parameters into a complete data set;

[0029] Use KNN density estimation to calculate the LOF value of each parameter component of the complete data set, and determine the local abnormality degree of each parameter component;

[0030] Set an abnormality threshold, screen out the abnormal components with LOF values greater than the abnormality threshold, and construct an abnormal component set;

[0031] Analyze the set of abnormal components to find the key abnormal components that cause the abnormality of the experimental data.

[0032] The present invention also proposes a method for generating a visualization chart of microbial data based on model parameters, including:

[0033] Obtain the historical experimental data of microbial culture, and use the PCA method to reduce the dimension of the historical experimental data to obtain the reduced-dimensional experimental data; the historical experimental data includes historical experimental parameters and statistical microbial quantities;

[0034] Model the reduced-dimensional experimental data through a weighted LSTM model to obtain a pre-trained weighted LSTM model, record the test microbial quantity generated by the test set, and compare it with the statistical microbial quantity to generate a comparison visualization image;

[0035] Record the initial microbial quantity, predict the microbial quantity and collect the microbial quantity every first time interval to obtain the predicted microbial quantity and the real-time microbial quantity, calculate the real-time error rate between the two, and generate an error visualization image;

[0036] Judge the relationship between the real-time error rate and the error threshold. If the real-time error rate is greater than the error threshold, mark the calculation time of the real-time error rate as an abnormal point, and mark the abnormal point in the error visualization image;

[0037] Obtain the abnormal experimental parameters at the abnormal point, calculate the LOF value of each component of the abnormal experimental parameters through the local outlier factor algorithm, obtain the key abnormal components of the abnormal experimental parameters, and output the visualization image of the key abnormal components.

[0038] Further, using the PCA method to reduce the dimension of the historical experimental data includes:

[0039] Fill in the missing values, remove the outliers and perform standardization processing on the historical experimental parameters of the historical experimental data to obtain the standardized experimental parameters;

[0040] Calculate the covariance matrix of the standardized experimental parameters, solve the eigenvalues and the corresponding eigenvectors of the covariance matrix, and obtain the principal component change direction;

[0041] Select the number of principal component spaces according to the cumulative variance contribution rate, and project the standardized experimental parameters into the selected principal component space to obtain the principal component parameters.

[0042] Further, the reduced-dimensional experimental data includes principal component parameters and statistical microbial quantities, and the principal component parameters include culture environment parameters, microbial characteristic parameters and culture process parameters.

[0043] Further, the weighted LSTM model includes:

[0044] A data preprocessing module for cleaning and normalizing the input data;

[0045] An LSTM module containing three LSTM sub-modules, which are respectively used to process the culture environment parameters, microbial characteristic parameters, and culture process parameters;

[0046] A weight assignment module for assigning weights to the outputs of each LSTM sub-module;

[0047] A weighted fusion module: summing the outputs of different LSTM sub-modules with weights to obtain a weighted hidden state;

[0048] An output module for predicting the weighted hidden state and outputting the predicted microbial biomass;

[0049] The formula for the weighted hidden state is:

[0050] h final = ω env * h env + ω micro * h micro + ω process * h process ;

[0051] where h final represents the weighted hidden state, h env represents the environmental hidden state, h micro represents the characteristic hidden state, h process represents the process hidden state, ω env represents the environmental weight coefficient, ω micro represents the characteristic weight coefficient, ω process represents the process weight coefficient.

[0052] Further, the abnormal components for obtaining abnormal experimental parameters include:

[0053] Obtain abnormal experimental parameters, and combine the historical experimental parameters and the abnormal experimental parameters into a complete data set;

[0054] Use KNN density estimation to calculate the LOF value of each parameter component of the complete data set, and determine the local abnormality degree of each parameter component;

[0055] Set an abnormality threshold, filter out the abnormal components with LOF values greater than the abnormality threshold, and construct an abnormal component set;

[0056] Analyze the abnormal component set to find the key abnormal components that cause the abnormality of the experimental data.

[0057] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0058] 1. The historical experimental data of microbial culture is optimized through PCA dimensionality reduction, improving data quality and calculation efficiency. First, missing value filling and outlier removal ensure data integrity, and normalization eliminates the difference in the dimension of different parameters, enhancing data consistency. Second, covariance matrix analysis combined with eigenvalue and eigenvector calculations can accurately identify the main component change direction of experimental parameters and extract the most representative feature information. Finally, the optimal principal component dimension is selected through the cumulative variance contribution rate, realizing data dimensionality reduction while retaining key information and reducing redundancy.

[0059] 2. The dimensionality-reduced experimental data is modeled through a weighted LSTM model, improving the prediction accuracy and data utilization efficiency of the microbial culture process. Through data preprocessing to ensure input quality, the LSTM sub-module extracts the key features of the culture environment, microbial characteristics, and culture process respectively, and the weight allocation module dynamically adjusts the contributions of each module to optimize information fusion. Finally, the model accurately predicts the number of microorganisms, and visually displays the comparison between the prediction results and the actual statistical data through visual images, providing efficient and accurate data support for the optimization and scientific decision-making of the microbial culture process.

[0060] 3. The key abnormal components of abnormal experimental parameters are accurately identified through the local outlier factor algorithm, improving the accuracy of anomaly detection. First, a complete data set is constructed in combination with historical experimental data to ensure the comprehensiveness of anomaly analysis. Then, the LOF values of each parameter component are calculated using KNN density estimation to quantify its local anomaly degree. By setting an anomaly threshold, the abnormal components are screened out to construct an abnormal component set, accurately locking the key factors leading to experimental anomalies. Finally, the key abnormal components are visually displayed through visual images, providing a scientific basis for experimental optimization, anomaly tracing, and data correction. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 It is a schematic structural diagram of a microbial data visualization chart generation system based on model parameters proposed by the present invention;

[0062] Figure 2 It is a schematic structural diagram of the weighted LSTM model proposed by the present invention;

[0063] Figure 3 It is a flowchart of a method for generating a microbial data visualization chart based on model parameters proposed by the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0064] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0065] Embodiment 1

[0066] A microbial data visualization chart generation system based on model parameters, as Figure 1 shown, includes:

[0067] A principal component parameter acquisition module, configured to acquire historical experimental data of microbial culture, perform dimensionality reduction on the historical experimental data using the PCA method, and obtain dimensionality-reduced experimental data; the historical experimental data includes historical experimental parameters and statistical microbial quantities.

[0068] Specifically, the historical experimental data is time series data, each time series interval is the first time interval, and the historical experimental data of each time series includes historical experimental parameters and statistical microbial quantities; in this embodiment, the first time interval is 6 hours, and the historical experimental parameters include parameters such as temperature, pH value, dissolved oxygen concentration, humidity, light condition, microbial activity, and metabolite concentration.

[0069] Further, performing dimensionality reduction on the historical experimental data using the PCA method includes:

[0070] Performing missing value filling, outlier removal, and standardization processing on the historical experimental parameters of the historical experimental data to obtain standardized experimental parameters;

[0071] Calculating the covariance matrix of the standardized experimental parameters, solving the eigenvalues and corresponding eigenvectors of the covariance matrix, and obtaining the principal component change direction;

[0072] Selecting the number of principal component spaces according to the cumulative variance contribution rate, and projecting the standardized experimental parameters into the selected principal component space to obtain principal component parameters.

[0073] Specifically, the missing values in all historical experimental parameters are filled using the mean filling method, the outliers of each parameter are detected through the Z-score algorithm, and the values exceeding 3 times the standard deviation are regarded as outliers and removed. All historical experimental parameters are standardized to have zero mean and unit variance, which is convenient for subsequent PCA dimensionality reduction processing.

[0074] Dimensionality reduction of historical experimental data on microbial culture through the PCA method can improve data quality, reduce noise, eliminate the influence of dimension through standardization processing, and enhance the stability and computational efficiency of the model. Covariance matrix analysis ensures that key features are retained, and the cumulative variance contribution rate optimizes the selection of principal components, enabling the data after dimensionality reduction to reduce redundancy while maintaining the main information and avoiding overfitting.

[0075] Further, the dimensionality-reduced experimental data includes principal component parameters and statistical microbial quantities, and the principal component parameters include culture environment parameters, microbial characteristic parameters, and culture process parameters.

[0076] Specifically, the dimensionality-reduced experimental data is obtained by dimensionality reduction of historical experimental data, where the historical experimental parameters as independent variables are reduced to principal component parameters, and the statistical microbial quantities as dependent variables remain unchanged. Part of the dimensionality-reduced experimental data under a certain time series is shown in Table 1.

[0077] Table 1 Part of the dimensionality-reduced experimental data under a certain time series

[0078] Experiment number Cultivation environment parameters Microbial characteristic parameters Cultivation process parameters Count the number of microorganisms 1 0.32 0.45 -0.18 500 2 0.38 0.52 -0.22 510 3 0.25 0.40 -0.10 480 4 0.59 0.65 -0.25 520 5 0.54 0.63 -0.20 530

[0079] Through the dimensionality-reduced experimental data, redundant information can be reduced, computational efficiency can be improved, while key features are retained, enhancing the interpretability of the data. The division of principal component parameters enables refined representation of culture environment, microbial characteristics, and culture process factors, helping to identify the core variables affecting microbial growth.

[0080] The comparison data visualization module is used to model the dimensionality-reduced experimental data through a weighted LSTM model to obtain a pre-trained weighted LSTM model, record the test microbial quantity generated by the test set, and compare it with the statistical microbial quantity to generate a comparison visualization image;

[0081] Further, the weighted LSTM model is as Figure 2 shown and includes:

[0082] The data preprocessing module is used to clean and standardize the input data;

[0083] The LSTM module contains three LSTM sub-modules, which are respectively used to process culture environment parameters, microbial characteristic parameters, and culture process parameters;

[0084] The weight assignment module is used to assign weights to the outputs of each LSTM sub-module;

[0085] The weighted fusion module: sums the outputs of different LSTM sub-modules with weights to obtain a weighted hidden state;

[0086] The output module is used to predict the weighted hidden state and output the predicted microbial quantity;

[0087] The weighted hidden state calculation formula is as follows:

[0088] h final = ω env * h env + ω micro * h micro + ω process * h process ;

[0089] Among them, h final represents the weighted hidden state, h env represents the environmental hidden state, h micro represents the characteristic hidden state, h process represents the process hidden state, ω env represents the environmental weight coefficient, ω micro represents the characteristic weight coefficient, ω process represents the process weight coefficient.

[0090] Specifically, the three LSTM sub-modules of the LSTM module respectively output h env , h micro and h process , and these outputs can be expressed as:

[0091] h env = LSTM(X env , θ env );

[0092] h micro = LSTM(X micro , θ micro );

[0093] h process = LSTM(X process , θ process );

[0094] Among them, LSTM() represents the LSTM sub-module, X env represents the cultivation environment parameter, X micro represents the microbial characteristic parameter, X process represents the cultivation process parameter, θ env , θ micro and θ process respectively represent the weights and biases of the LSTM sub-module; the weight assignment module assigns weight coefficients to the outputs of each LSTM sub-module, and the calculation formula of the weight coefficient can be expressed as:

[0095]

[0096] Among them, exp() represents the exponential function, and f() represents the scoring function, which can be a simple feed-forward neural network or other learning mechanisms. In this embodiment, a weighted average function calculated by an attention mechanism is adopted; the final output module outputs the prediction result, and this process can be expressed as:

[0097]

[0098] Among them, W f represents the weight matrix of the output module, and b f represents the bias term of the output module.

[0099] By using a weighted LSTM model to model the dimensionality-reduced experimental data, not only can the temporal features of the key factors in the microbial culture process be effectively extracted, but also the culture environment, microbial characteristics, and culture process data can be processed by different LSTM sub-modules respectively, improving the pertinence and accuracy of feature extraction. The weight allocation module dynamically adjusts the influence weights of different sub-modules, enabling the model to adaptively focus on the most important influencing factors, thereby improving the prediction accuracy and robustness. The weighted fusion module further optimizes the information integration, making the contributions of different factors more reasonable. Finally, the microbial quantity is accurately predicted through the output module, and a comparative visualization image is generated to intuitively display the prediction effect.

[0100] The error visualization module is used to record the initial microbial quantity, perform microbial quantity prediction and microbial quantity collection every first time interval, obtain the predicted microbial quantity and the real-time microbial quantity, calculate the real-time error rate between the two, and generate an error visualization image;

[0101] Specifically, the calculation formula of the real-time error rate is:

[0102]

[0103] Among them, E represents the real-time error rate, y represents the real-time microbial quantity, and y p represents the predicted microbial quantity.

[0104] The anomaly detection module is used to judge the relationship between the real-time error rate and the error threshold. If the real-time error rate is greater than the error threshold, mark the calculation time of the real-time error rate as an anomaly point, and mark the anomaly point in the error visualization image; in this embodiment, the error threshold is 0.2.

[0105] The abnormal data visualization module is used to obtain the abnormal experimental parameters at the anomaly point, calculate the LOF value of each component of the abnormal experimental parameters through the local outlier factor algorithm, obtain the key abnormal components of the abnormal experimental parameters, and output the visualization image of the key abnormal components.

[0106] Further, the abnormal components for obtaining abnormal experimental parameters include:

[0107] Obtain abnormal experimental parameters, and merge the historical experimental parameters and the abnormal experimental parameters into a complete data set;

[0108] Use KNN density estimation to calculate the LOF value of each parameter component of the complete data set, and determine the local abnormality degree of each parameter component;

[0109] Set an abnormal threshold, screen out the abnormal components with LOF values greater than the abnormal threshold, and construct an abnormal component set;

[0110] Analyze the abnormal component set to find the key abnormal components that cause the abnormality of the experimental data.

[0111] Specifically, obtain the abnormal experimental parameters at the abnormal point, that is, the abnormal experimental data when the abnormality occurs. The abnormal experimental parameters are a 1×N-dimensional vector, where N represents N components of the experimental parameters, including temperature, pH value, dissolved oxygen concentration, humidity, light condition, microbial activity, and metabolite concentration, etc. The historical experimental parameters are an M×N-dimensional vector, where M represents the number of groups of experiments. Therefore, the complete data set X is a (M + 1)×N-dimensional vector, which can be expressed as X = {x1, x2,..., x N}; The calculation formula for the LOF value of each parameter component can be expressed as:

[0112]

[0113] where LOF represents the LOF value of the parameter component, k represents the number of the nearest neighbors in the KNN algorithm, N k (x i ) represents the k nearest neighbors of point x i , and lrd(x i ) represents the local reachability density of point x i , which can be expressed as:

[0114]

[0115] where rd(x i , x j ) represents the reachable distance between point x i and point x j , which can be expressed as:

[0116] rd(x i , x j ) = max(d(x i , x j ), k_dist(x j ));

[0117] Among them, max() represents the maximum value function, and d(x i , x j ) represents the Euclidean distance between two points, and k_dist(x j ) represents the distance from point x j to its nearest neighbor; the anomaly threshold is set to 1.5, that is, when LOF > 1.5, the point is considered an abnormal component, and an abnormal component set is constructed; the method for analyzing the abnormal component set can directly take the maximum value in the abnormal component set as the key abnormal component, or use the independent component analysis method to analyze the relationship between each abnormal component and then determine the key abnormal component. In this embodiment, the maximum value in the abnormal component set is directly taken as the key abnormal component. Table 2 is the LOF data table of some experimental data.

[0118] By finely analyzing each component of the abnormal experimental parameters through the local outlier factor algorithm, the key factors leading to experimental anomalies can be effectively identified. Combining historical experimental parameters to construct a complete data set and using KNN density estimation to calculate the LOF value make the anomaly detection more robust and reliable. By setting the anomaly threshold, the most abnormal components can be accurately screened out, an abnormal component set can be constructed, and thus the core influencing factors of experimental anomalies can be clarified.

[0119] Table 2 LOF data table of some experimental data

[0120] Experiment number Parameter component 1 Parameter component 2 Parameter component 3 LOF1 LOF2 LOF3 1 25.4 3.2 45.6 0.75 1.20 0.80 2 30.2 2.7 47.9 1.45 0.98 1.05 3 32.1 3.9 50.8 1.65 1.30 1.55 4 26.5 3.1 49.0 0.90 1.10 0.92

[0121] This system combines PCA dimensionality reduction and weighted LSTM modeling, improving the processing efficiency and prediction accuracy of microbial culture experimental data. Through PCA dimensionality reduction, the key features of the culture environment, microbial characteristics, and culture process are extracted, redundant information is reduced, and the model has stronger generalization ability. Comparing the prediction based on the weighted LSTM model with the actual statistical microbial quantity visualizes the experimental trend and helps optimize the culture conditions. The real-time error calculation and anomaly detection mechanism can dynamically monitor the microbial growth situation, accurately identify the abnormal time points, and deeply analyze the abnormal experimental parameters through the LOF algorithm to find the key abnormal components, and finally generate a visualized image to intuitively display the source and influencing factors of the anomalies.

[0122] Embodiment 2

[0123] A method for generating a visualized chart of microbial data based on model parameters, as Figure 3 shown, includes:

[0124] Obtain historical experimental data of microbial culture, and use the PCA method to reduce the dimension of the historical experimental data to obtain dimension-reduced experimental data; the historical experimental data includes historical experimental parameters and statistical microbial quantity;

[0125] Model the dimensionality-reduced experimental data through a weighted LSTM model to obtain a pre-trained weighted LSTM model, record the test microbial quantity generated by the test set, and compare it with the statistical microbial quantity to generate a comparative visualization image;

[0126] Record the initial microbial quantity, perform microbial quantity prediction and microbial quantity collection every first time interval to obtain the predicted microbial quantity and the real-time microbial quantity, calculate the real-time error rate between the two, and generate an error visualization image;

[0127] Judge the relationship between the real-time error rate and the error threshold. If the real-time error rate is greater than the error threshold, mark the calculation time of the real-time error rate as an abnormal point, and mark the abnormal point in the error visualization image;

[0128] Obtain the abnormal experimental parameters at the abnormal point, calculate the LOF value of each component of the abnormal experimental parameters through the local outlier factor algorithm, obtain the key abnormal components of the abnormal experimental parameters, and output the visualization image of the key abnormal components.

[0129] Further, using the PCA method to reduce the dimension of the historical experimental data includes:

[0130] Fill in the missing values, remove the outliers and perform standardization processing on the historical experimental parameters of the historical experimental data to obtain standardized experimental parameters;

[0131] Calculate the covariance matrix of the standardized experimental parameters, solve the eigenvalues and the corresponding eigenvectors of the covariance matrix, and obtain the main component change direction;

[0132] Select the number of main component spaces according to the cumulative variance contribution rate, and project the standardized experimental parameters into the selected main component space to obtain the main component parameters.

[0133] Further, the dimensionality-reduced experimental data includes main component parameters and statistical microbial quantities, and the main component parameters include culture environment parameters, microbial characteristic parameters and culture process parameters.

[0134] Further, the weighted LSTM model includes:

[0135] A data preprocessing module for cleaning and standardizing the input data;

[0136] An LSTM module containing three LSTM sub-modules, which are respectively used to process the culture environment parameters, microbial characteristic parameters and culture process parameters;

[0137] A weight allocation module for allocating weights to the outputs of each LSTM sub-module;

[0138] Weighted Fusion Module: The outputs of different LSTM sub-modules are weighted and summed to obtain a weighted hidden state;

[0139] Output Module, which is used to predict the weighted hidden state and output the predicted microbial biomass;

[0140] The calculation formula for the weighted hidden state is:

[0141] h final = ω env * h env + ω micro * h micro + ω process * h process ;

[0142] Wherein, h final represents the weighted hidden state, h env represents the environmental hidden state, h micro represents the characteristic hidden state, h process represents the process hidden state, ω env represents the environmental weight coefficient, ω micro represents the characteristic weight coefficient, ω process represents the process weight coefficient.

[0143] Furthermore, the abnormal components for obtaining abnormal experimental parameters include:

[0144] Obtain abnormal experimental parameters, and combine the historical experimental parameters and the abnormal experimental parameters into a complete data set;

[0145] Use KNN density estimation to calculate the LOF value of each parameter component of the complete data set, and determine the local abnormality degree of each parameter component;

[0146] Set an abnormality threshold, filter out the abnormal components with LOF values greater than the abnormality threshold, and construct an abnormal component set;

[0147] Analyze the abnormal component set to find the key abnormal components that cause the abnormality of the experimental data.

[0148] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A microbial data visualization chart generation system based on model parameters, characterized in that, Including: A main component parameter acquisition module, which is used to acquire historical experimental data of microbial culture, reduce the dimension of the historical experimental data by using the PCA method, and obtain reduced-dimensional experimental data; the historical experimental data includes historical experimental parameters and statistical microbial quantities; A comparison data visualization module, which is used to model the reduced-dimensional experimental data through a weighted LSTM model to obtain a pre-trained weighted LSTM model, record the test microbial quantity generated by the test set, compare it with the statistical microbial quantity, and generate a comparison visualization image; An error visualization module, which is used to record the initial microbial quantity, predict the microbial quantity and collect the microbial quantity every first time interval to obtain the predicted microbial quantity and the real-time microbial quantity, calculate the real-time error rate between the two, and generate an error visualization image; An anomaly detection module, which is used to judge the relationship between the real-time error rate and the error threshold. If the real-time error rate is greater than the error threshold, mark the calculation time of the real-time error rate as an anomaly point, and mark the anomaly point in the error visualization image; An abnormal data visualization module, which is used to obtain the abnormal experimental parameters at the anomaly point, calculate the LOF value of each component of the abnormal experimental parameters through the local outlier factor algorithm, obtain the key abnormal components of the abnormal experimental parameters, and output the visualization image of the key abnormal components.

2. The microbial data visualization chart generation system based on model parameters according to claim 1, wherein Using the PCA method to reduce the dimension of the historical experimental data includes: Performing missing value filling, outlier removal and standardization processing on the historical experimental parameters of the historical experimental data to obtain standardized experimental parameters; Calculating the covariance matrix of the standardized experimental parameters, solving the eigenvalues and corresponding eigenvectors of the covariance matrix, and obtaining the main component change direction; Selecting the number of main component spaces according to the cumulative variance contribution rate, and projecting the standardized experimental parameters into the selected main component space to obtain main component parameters.

3. The microbial data visualization chart generation system based on model parameters according to claim 1, characterized in that, The reduced-dimensional experimental data includes main component parameters and statistical microbial quantities, and the main component parameters include culture environment parameters, microbial characteristic parameters and culture process parameters.

4. The microbial data visualization chart generation system based on model parameters according to claim 1, wherein The weighted LSTM model includes: A data preprocessing module, which is used to clean and standardize the input data; An LSTM module, which contains three LSTM sub-modules, which are respectively used to process culture environment parameters, microbial characteristic parameters and culture process parameters; A weight allocation module, which is used to allocate weights to the outputs of each LSTM sub-module; A weighted fusion module: performing weighted summation on the outputs of different LSTM sub-modules to obtain a weighted hidden state; An output module, which is used to predict the weighted hidden state and output the predicted microbial quantity; The formula for the weighted hidden state is: h final = ω env * h env + ω micro * h micro + ω process * h process ; Among them, h final represents the weighted hidden state, h env represents the environmental hidden state, h micro represents the feature hidden state, h process represents the process hidden state, ω env represents the environmental weight coefficient, ω micro represents the feature weight coefficient, ω process represents the process weight coefficient.

5. The microbial data visualization chart generation system based on model parameters according to claim 1, wherein Obtaining the abnormal components of the abnormal experimental parameters includes: Obtaining the abnormal experimental parameters, and combining the historical experimental parameters and the abnormal experimental parameters into a complete data set; Calculating the LOF value of each parameter component of the complete data set by using KNN density estimation to determine the local outlier degree of each parameter component; Setting an outlier threshold, screening out the outlier components with LOF values greater than the outlier threshold, and constructing an outlier component set; Analyze the set of abnormal components to find the key abnormal components that cause the abnormality of the experimental data.

6. A method for generating a visualization chart of microbial data based on model parameters, characterized in that, Including: Obtain the historical experimental data of microbial culture, and use the PCA method to reduce the dimension of the historical experimental data to obtain the dimension-reduced experimental data; the historical experimental data includes historical experimental parameters and statistical microbial quantities; Model the dimension-reduced experimental data through a weighted LSTM model to obtain a pre-trained weighted LSTM model, record the test microbial quantity generated by the test set, and compare it with the statistical microbial quantity to generate a comparison visualization image; Record the initial microbial quantity, perform microbial quantity prediction and microbial quantity collection at every first time interval to obtain the predicted microbial quantity and the real-time microbial quantity, calculate the real-time error rate between the two, and generate an error visualization image; Judge the relationship between the real-time error rate and the error threshold. If the real-time error rate is greater than the error threshold, mark the calculation time of the real-time error rate as an abnormal point, and mark the abnormal point in the error visualization image; Obtain the abnormal experimental parameters at the abnormal point, calculate the LOF value of each component of the abnormal experimental parameters through the local outlier factor algorithm, obtain the key abnormal components of the abnormal experimental parameters, and output the visualization image of the key abnormal components.

7. The method for generating a visualization chart of microbial data based on model parameters according to claim 6, wherein Using the PCA method to reduce the dimension of the historical experimental data includes: Fill in the missing values, remove the outliers and standardize the historical experimental parameters of the historical experimental data to obtain the standardized experimental parameters; Calculate the covariance matrix of the standardized experimental parameters, solve the eigenvalues and the corresponding eigenvectors of the covariance matrix, and obtain the main component change direction; Select the number of main component spaces according to the cumulative variance contribution rate, and project the standardized experimental parameters into the selected main component space to obtain the main component parameters.

8. The method for generating a visualization chart of microbial data based on model parameters according to claim 6, characterized in that, The dimension-reduced experimental data includes main component parameters and statistical microbial quantities, and the main component parameters include culture environment parameters, microbial characteristic parameters and culture process parameters.

9. The method for generating a visualization chart of microbial data based on model parameters according to claim 6, characterized in that, The weighted LSTM model includes: A data preprocessing module for cleaning and standardizing the input data; An LSTM module containing three LSTM sub-modules, which are respectively used to process the culture environment parameters, microbial characteristic parameters and culture process parameters; A weight allocation module for allocating weights to the outputs of each LSTM sub-module; A weighted fusion module: sum the outputs of different LSTM sub-modules with weights to obtain a weighted hidden state; An output module for predicting the weighted hidden state and outputting the predicted microbial quantity; The formula for the weighted hidden state is: h final = ω env * h env + ω micro * h micro + ω process * h process ; where, h final represents the weighted hidden state, h env represents the environmental hidden state, h micro represents the feature hidden state, h process represents the process hidden state, ω env represents the environmental weight coefficient, ω micro represents the feature weight coefficient, ω process represents the process weight coefficient.

10. The method for generating a visualized chart of microbial data based on model parameters according to claim 6, wherein Obtaining the abnormal components of the abnormal experimental parameters includes: Obtain the abnormal experimental parameters, and merge the historical experimental parameters and the abnormal experimental parameters into a complete data set; Use KNN density estimation to calculate the LOF value of each parameter component of the complete data set to determine the local outlier degree of each parameter component; Set an abnormal threshold, screen out the abnormal components whose LOF values are greater than the abnormal threshold, and construct a set of abnormal components; Analyze the set of abnormal components to find the key abnormal components that cause the abnormality of the experimental data.

Images