Sample analysis method and system based on multi-dimensional gas chromatograph data fusion algorithm

Abstract

The invention discloses a sample analysis method and system based on a multi-dimensional gas chromatograph data fusion algorithm, and belongs to the technical field of sample analys.The method comprises the steps that 1, information of a multi-dimensional gas chromatograph is collected, and the data fusion algorithm is set according to the information of the multi-dimensional gas chromatograph; 2, identifying the detection sample, and detecting the detection sample through a multi-dimensional gas chromatograph to obtain corresponding sample detection data; 3, determining debugging parameters of a data fusion algorithm according to the sample detection data, and debugging the data fusion algorithm according to the debugging parameters; 4, performing fusion processing on the sample detection data through a debugged data fusion algorithm to obtain fused detection data; and 5, obtaining a sample analysis target of the user, and processing the fusion detection data according to the sample analysis target to obtain a sample analysis result.

Description

Technical Field

[0001] The present invention belongs to the technical field of sample analysis, and in particular relates to a sample analysis method and system based on a multi-dimensional gas chromatograph data fusion algorithm. Background Art

[0002] Gas chromatography is a precision technology widely used in the field of chemical analysis, especially in the analysis of complex samples. However, with the development of science and technology and the increase in sample complexity, traditional single-dimensional gas chromatography has been unable to meet the analysis needs. In order to overcome this challenge, multidimensional gas chromatography came into being.

[0003] The core of multidimensional gas chromatography is to use two or more analytical methods that have their own advantages but are completely different to jointly analyze compounds. Among them, the most well-known multidimensional analysis technology is gas chromatography-mass spectrometry (GC-MS), which combines the separation ability of gas chromatography and the detection ability of mass spectrometry, significantly improving the accuracy and sensitivity of the analysis results. However, traditional multidimensional gas chromatography also has certain challenges in data processing. Since data from different dimensions and different detectors may have differences in format and standards, data inconsistency may be encountered during data fusion. At the same time, with the increase in data volume, the complexity of data processing and analysis is also increasing, which puts higher requirements on data processing algorithms and computing power.

[0004] In order to solve the above problems, the present invention provides a sample analysis method and system based on a multi-dimensional gas chromatograph data fusion algorithm. Summary of the invention

[0005] In order to solve the problems existing in the above scheme, the present invention provides a sample analysis method and system based on a multi-dimensional gas chromatograph data fusion algorithm.

[0006] The purpose of the present invention can be achieved through the following technical solutions: A sample analysis method based on a multidimensional gas chromatograph data fusion algorithm, the method comprising: Step 1: Collect multi-dimensional gas chromatograph information, and set a data fusion algorithm according to the multi-dimensional gas chromatograph information; Furthermore, the method for setting the data fusion algorithm includes: Determining a sample detection range of the multidimensional gas chromatograph according to the multidimensional gas chromatograph information; Determine the potential fusion data according to the multi-dimensional gas chromatograph information and the sample detection range; perform a search according to the potential fusion data to obtain a plurality of fusion algorithms to be selected; The candidate fusion algorithms are screened according to the sample detection range to determine the data fusion algorithm to be applied.

[0007] Furthermore, the method for screening the fusion algorithm to be selected according to the sample detection range includes: Identify the sample types corresponding to the sample detection range, and count the detection proportions of the sample types; Performing a detection simulation on the sample type by using the candidate fusion algorithm to obtain a fusion effect value of data fusion performed by the candidate fusion algorithm on the sample type; The sample type is marked as i, i=1, 2, ..., n, and n is the number of sample types; The obtained fusion effect value is marked as RH i ; Mark the detection specific gravity of the sample type as δ i ; The screening value of the corresponding candidate fusion algorithm is calculated according to the screening formula. The screening formula is: ; Where: SW is the screening value; The candidate fusion algorithm with the largest screening value is selected as the applied data fusion algorithm.

[0008] Furthermore, the method of simulating the detection of sample types by using the selected fusion algorithm includes: Acquire fusion materials corresponding to the sample types, perform fusion processing on the fusion materials through the selected fusion algorithm, and obtain fusion result data; Preset fusion effect evaluation indicators and indicator sub-areas, and set corresponding reference standards and standard discount methods according to the fusion effect evaluation indicators; According to the fusion effect evaluation index, the fusion result data is collected to obtain the index analysis data corresponding to the fusion effect evaluation index; according to the reference standard and the standard discount method, the index analysis data is analyzed to obtain the index score corresponding to the fusion effect evaluation index; The index scores corresponding to each fusion effect evaluation index are accumulated to obtain the fusion effect value.

[0009] Step 2: Identify the test sample, test the test sample through a multi-dimensional gas chromatograph, and obtain corresponding sample test data; Step 3: determining the debugging parameters of the data fusion algorithm according to the sample detection data, and debugging the data fusion algorithm according to the debugging parameters; Furthermore, the method for determining the debugging parameters of the data fusion algorithm includes: The platform sets a parameter representative curve, wherein the horizontal axis of the parameter representative curve is the serial number value and the vertical axis is the parameter representative value; each serial number value corresponds to a sample feature set data; each parameter representative value corresponds to a debugging reference parameter; Identify the sample detection data of the detection sample, identify the sample feature set data corresponding to the sample detection data, match the corresponding debugging reference parameters for the sample feature set data according to the parameter representative curve, and mark them as debugging parameters.

[0010] Furthermore, the method for setting the parameter representative curve includes: Acquire historical test data of the sample type, determine historical sample test data of the sample type according to the historical test data, and mark them as sample material data; and determine a historical debugging parameter set corresponding to the sample material data according to the historical test data; The platform presets each sample feature, performs feature processing on the sample material data according to the sample feature, and obtains sample feature set data corresponding to the sample material data; Identify the historical debugging parameter set corresponding to the sample feature set data, evaluate the fusion effect value corresponding to each historical debugging parameter in the historical debugging parameter set; generate a fusion effect curve according to the fusion effect value and the corresponding historical debugging parameter; determine the debugging reference parameter of the sample feature set data according to the fusion effect curve; Sorting the sample feature set data corresponding to the sample types to obtain a first sequence; setting corresponding debugging reference parameters for the corresponding sample feature set data in the first sequence; Setting corresponding sequence number values ​​for each sample feature set data according to the first sequence; Mark the debugging reference parameter with a serial number value of 1 as a benchmark parameter, and set corresponding parameter representative values ​​for each debugging reference parameter according to the benchmark parameter; A parameter representative curve is generated according to the serial number value and the parameter representative value, with the horizontal axis being the serial number value and the vertical axis being the parameter representative value.

[0011] Furthermore, the setting method of the first sequence includes: Step SA1: Determine the first ranked sample feature set data in a preset manner to form an initial sequence, and determine the benchmark data based on the initial sequence, where the benchmark data is the last ranked sample feature set data in the initial sequence; Step SA2: Calculate the similarity between the benchmark data and the remaining sample feature set data, add the sample feature set data with the highest similarity to the initial sequence, obtain a new initial sequence, and determine new benchmark data based on the new initial sequence; Step SA3: Loop step SA2 until there is no remaining sample feature set data, and mark the initial sequence as the first sequence.

[0012] Furthermore, when the debugging result of debugging the data fusion algorithm according to the debugging parameters does not meet the user's requirements, the platform performs remote control debugging.

[0013] Furthermore, the method for remote control debugging by the platform includes: When the user has a need for remote control, the user opens the data management authority and sends remote control information to the platform; when the platform receives the remote control information, it collects debugging and analysis data in real time after the data management authority is opened; The debugging analysis data is analyzed to determine debugging parameters, and the data fusion algorithm is debugged according to the debugging parameters.

[0014] Step 4: Fusing the sample test data through the debugged data fusion algorithm to obtain fused test data; Step 5: Obtain the user's sample analysis target, process the fusion detection data according to the sample analysis target, and obtain the sample analysis result.

[0015] A sample analysis system based on a multi-dimensional gas chromatograph data fusion algorithm, comprising an equipment analysis module, a detection module, an algorithm debugging module and a data analysis module; The equipment analysis module is used to analyze the multi-dimensional gas chromatograph and set the data fusion algorithm for application; The detection module is used to detect the detection sample and obtain corresponding sample detection data; The algorithm debugging module is used to debug the data fusion algorithm; The data analysis module is used to perform detection analysis, fuse the sample detection data through the debugged data fusion algorithm to obtain fused detection data; obtain the user's sample analysis target, process the fused detection data according to the sample analysis target, and obtain the sample analysis result.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The data fusion algorithm based on multidimensional gas chromatograph provided by the present invention can effectively solve the data format and standard inconsistency problems encountered in data processing of traditional multidimensional gas chromatography. The algorithm can automatically identify and match data generated by different dimensions and different detectors, ensuring the consistency and accuracy of the data during the fusion process, thereby greatly improving the reliability and accuracy of the analysis results. By solving the bottleneck problem in data processing, the present invention greatly promotes the widespread application of multidimensional gas chromatography-mass spectrometry (GC-MS) and other multidimensional analysis technologies in the fields of chemistry, environmental science, food safety, and drug development. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0019] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] like Figure 1 As shown, a sample analysis method based on a multidimensional gas chromatograph data fusion algorithm comprises: Step 1: Collect multi-dimensional gas chromatograph information and set a corresponding data fusion algorithm according to the multi-dimensional gas chromatograph information.

[0021] The multi-dimensional gas chromatograph information is collected according to an actual multi-dimensional gas chromatograph.

[0022] The data fusion algorithm is used to fuse the sample detection data from different dimensions and different detectors.

[0023] In one embodiment, the data fusion algorithm may be set based on existing methods, such as based on the working experience and professional ability of professionals.

[0024] In one embodiment, a method for setting a data fusion algorithm according to multi-dimensional gas chromatograph information includes: Determine the sample types that the multi-dimensional gas chromatograph can detect and analyze based on the multi-dimensional gas chromatograph information, and integrate them into the sample detection range; According to the multi-dimensional gas chromatograph information and the sample detection range, various subsequent sample detection data are determined, and the potential fusion data is set, that is, the sample detection data that may need to be fused in the future; according to the potential fusion data, a search is performed to determine the various existing data fusion algorithms that can meet the potential fusion data fusion processing and can be applied, and they are marked as candidate fusion algorithms, that is, the candidate fusion algorithm is an existing data fusion algorithm, or a data fusion algorithm set by the platform, etc., which can be used in conjunction with the multi-dimensional gas chromatograph.

[0025] The fusion algorithms to be selected are screened according to the sample detection range, and the data fusion algorithm to be applied is determined.

[0026] In one embodiment, the fusion algorithms to be selected are screened according to the sample detection range, and the screening can be performed based on existing methods.

[0027] In one embodiment, the method for screening the fusion algorithms to be selected according to the sample detection range includes: Identify the sample types corresponding to the sample detection range, and determine the detection proportion of different sample types through historical detection data or prediction, that is, the proportion of the sample type in the subsequent detection of all sample types, and compare them according to their corresponding detection quantity; Through the candidate fusion algorithm, each sample type is tested and simulated, and the fusion effect value of the data fusion of the candidate fusion algorithm on the sample type is obtained. The historical sample test data corresponding to the sample type can be directly applied as the fusion material, and the candidate fusion algorithm is applied to perform fusion processing to determine the corresponding fusion result data, and the data with the best effect is selected, that is, the data with the best effect is selected from the data obtained by multiple simulations as the fusion result data; and then the corresponding fusion effect value is evaluated according to the fusion result data; The sample type is marked as i, i=1, 2, ..., n, and n is the number of sample types; The obtained fusion effect value is marked as RH i ; Mark the detection specific gravity of the sample type as δ i ; The screening value of the corresponding candidate fusion algorithm is calculated according to the screening formula. The screening formula is: ; Where: SW is the screening value; The candidate fusion algorithm with the largest screening value is selected as the applied data fusion algorithm.

[0028] In one embodiment, the fusion effect value can be in percentage, and various fusion effect evaluation indicators and indicator partition intervals are preset, and the indicator partition interval is the fusion effect value range corresponding to the fusion effect evaluation indicator, such as [0, 20]; and the reference standard and standard discount method corresponding to the fusion effect evaluation indicator are set; the reference standard is the requirement that needs to be met for the full score of the fusion effect evaluation indicator, that is, to meet the reference standard, and its indicator is divided into the maximum value between the indicator partitions; the standard discount method is the indicator score that needs to be discounted to different degrees below the reference standard; the fusion effect evaluation indicator, reference standard, standard discount method and indicator partition interval are all preset according to actual evaluation needs; Obtain fusion result data obtained by performing detection simulation on corresponding sample types by the selected fusion algorithm; According to each fusion effect evaluation indicator, the fusion result data is collected to obtain the indicator analysis data corresponding to the fusion effect evaluation indicator; according to the corresponding reference standard and standard discount method, the indicator analysis data is analyzed to obtain the indicator score corresponding to the corresponding fusion effect evaluation indicator; The index scores corresponding to each fusion effect evaluation index are accumulated to obtain the fusion effect value.

[0029] Exemplary: Preset fusion effect evaluation indicators: Data integrity: Evaluate whether the fused data contains all the information of the original data without omission.

[0030] Reference standard: Data integrity reaches more than 95%.

[0031] Data consistency: Evaluate whether the fused data is consistent across different dimensions and detectors without any contradictions.

[0032] Reference standard: Data consistency error is within 5%.

[0033] Data accuracy: Evaluate whether the fused data accurately reflects the true situation of the original data without bias.

[0034] Reference standard: Data accuracy reaches more than 98%.

[0035] Data correlation: Evaluate whether there is a reasonable correlation between different dimensions of the fused data, which can reflect the characteristics of the sample.

[0036] Reference standard: Data relevance score is above 80 points (based on a certain relevance assessment model).

[0037] Data redundancy: Evaluate whether there is redundant information in the fused data, that is, repeated or invalid data.

[0038] Reference standard: Data redundancy is controlled within 5%.

[0039] Set the indicator partition interval and standard discount method corresponding to the evaluation indicator: According to the preset fusion effect evaluation indicators and their reference standards, set the corresponding score range for each indicator. For example: Data integrity: 95%-100% will get full marks (e.g. 20 points), and 2 points will be deducted for every 1% decrease.

[0040] Data consistency: If the error is within 5%, full marks (e.g. 20 points) will be awarded, and 4 points will be deducted for every additional 1% error.

[0041] Data accuracy: 98%-100% will get full marks (e.g. 20 points), and 2 points will be deducted for every 1% decrease.

[0042] Data relevance: Full marks will be awarded for scores above 80 (e.g. 20 points), and 4 points will be deducted for every 5 points lower.

[0043] Data redundancy: Full marks (e.g. 20 points) will be awarded if the data redundancy is within 5%, and 4 points will be deducted for every 1% increase in redundancy.

[0044] Perform matching analysis and accumulate scores: Data matching analysis: Compare and analyze the fused data with the preset reference standards to determine the actual score of each evaluation indicator.

[0045] Accumulated score: The actual score of each evaluation indicator is accumulated to obtain the total score of the fusion effect.

[0046] Example of fusion effect value evaluation: Assume that after data fusion processing, the following evaluation results are obtained: Data completeness: 98% (18 points); Data consistency: 3% error (20 points); Data accuracy: 99% (18 points); Data relevance: 85 points (out of 20 points); Data redundancy: 4% (20 points); The total score of the fusion effect is: 18+20+18+20+20=96 points.

[0047] In one embodiment, the fusion effect value can also be evaluated based on existing methods, such as establishing an intelligent evaluation model based on intelligent algorithms such as neural networks, and performing intelligent evaluation through the successfully trained intelligent evaluation model to obtain the fusion effect value.

[0048] Step 2: Identify the samples that need to be tested and analyzed, mark them as test samples, test the test samples through a multi-dimensional gas chromatograph, and obtain corresponding sample test data.

[0049] Step 3: Determine the debugging parameters of the data fusion algorithm based on the sample detection data, and debug the data fusion algorithm according to the debugging parameters.

[0050] In one embodiment, a method for determining debugging parameters of a data fusion algorithm includes: Obtain the sample detection range of the multi-dimensional gas chromatograph, and identify each sample type according to the sample detection range; Obtain historical test data of each sample type, determine various historical sample test data of the sample type based on the historical test data, and mark them as sample material data; and determine the debugging parameters that already exist corresponding to the sample material data based on the historical test data, and integrate them into a historical debugging parameter set; The platform presets the sample characteristics of the sample detection data for parameter debugging of the data fusion algorithm, such as signal strength, noise level, timestamp, spatial position, data quality index, data distribution characteristics, correlation analysis, etc.; performs feature processing on the sample material data according to the sample characteristics, obtains the sample feature set data corresponding to the sample material data, that is, composed of the sample feature data corresponding to each sample feature, and analyzes the corresponding sample feature data using the existing method; Identify the historical debugging parameter set corresponding to the sample feature set data, and evaluate the fusion effect value corresponding to each historical debugging parameter in the historical debugging parameter set; generate a fusion effect curve according to the fusion effect value and the corresponding historical debugging parameter, with the debugging parameter as the horizontal axis and the fusion effect value as the vertical axis; determine the historical debugging parameter with the best fusion effect corresponding to the sample feature set data according to the fusion effect curve, and mark it as a debugging reference parameter. The debugging reference parameter can be determined by directly identifying the fusion effect value, or the highest possible fusion effect value can be determined in combination with the predicted data, and then the debugging reference parameter can be determined; Sorting each sample feature set data corresponding to the sample type to obtain a first sequence; setting corresponding debugging reference parameters for the corresponding sample feature set data in the first sequence; According to the first sequence, a corresponding sequence number value, i.e., a corresponding sequence number, is set for each sample feature set data; Mark the debugging reference parameter with a serial number value of 1 as a benchmark parameter, and set corresponding parameter representative values ​​for each debugging reference parameter according to the benchmark parameter; Generate a parameter representative curve according to the serial number value and the parameter representative value, with the horizontal axis being the serial number value and the vertical axis being the parameter representative value; Identify the sample test data, identify the sample feature set data corresponding to the sample test data, match the corresponding debugging reference parameters for the sample feature set data according to the parameter representative curve, and mark them as debugging parameters; the matching can be done directly, or a simulation test can be performed after matching the corresponding positions to determine the best debugging parameters. According to the curve changes, the optional range of the best debugging parameters for the sample feature set data can be greatly narrowed, thereby reducing the number of debugging times.

[0051] In one embodiment, the sample feature set data corresponding to the sample type are sorted based on existing sorting methods, such as evaluation priority, sorting in order of test results from good to bad, sorting in order of difficulty, manual sorting, and other sorting methods.

[0052] In one embodiment, the sample feature set data corresponding to the sample type are sorted, and the sample feature set data can be sorted according to the differences corresponding to the sample feature set data. The smaller the difference, the closer the two are sorted. The sample feature set data ranked first can be determined by optional, difficulty, probability of occurrence, etc., and the weight coefficients of different sample features are subsequently preset to calculate the similarity between the remaining sample feature sets and the sample feature set data ranked first. The sample feature set data with the highest similarity is regarded as ranked second, and the similarity between the remaining sample feature sets and the sample feature set data ranked second is calculated, and so on, until all sorting is completed. The specific method is: Step SA1: Determine the first ranked sample feature set data in a preset manner to form an initial sequence, and determine the benchmark data based on the initial sequence, where the benchmark data is the last ranked sample feature set data in the initial sequence; Step SA2: Calculate the similarity between the benchmark data and the remaining sample feature set data, add the sample feature set data with the highest similarity to the initial sequence, obtain a new initial sequence, and determine new benchmark data based on the new initial sequence; Step SA3: Loop step SA2 until there is no remaining sample feature set data, and mark the initial sequence as the first sequence.

[0053] In one embodiment, a corresponding parameter representative value is set for each debugging reference parameter according to the benchmark parameter. When it is determined that the parameter representative value of the benchmark parameter is 1, the parameters are sorted step by step in the above manner, that is, the similarity between the remaining debugging reference parameters and the benchmark parameters is calculated, and the parameter representative value of the debugging reference parameter with the greatest similarity is marked as 2, and so on, and are marked as 3, 4, 5, etc.

[0054] In one embodiment, corresponding parameter representative values ​​are set for various debugging reference parameters according to the benchmark parameters, which can be determined based on existing methods.

[0055] In one embodiment, signal strength: In many sensor detection systems, signal strength is a key feature. It reflects the energy of the signal received by the sensor. For data fusion algorithms, signal strength can help determine the reliability and weight of the data. When debugging parameters, signal strength may be used to adjust the weight distribution between different data sources in the fusion algorithm.

[0056] Noise level: Noise is the unwanted random variation in the data, which may come from the sensor itself, environmental interference, or errors in the data transmission process. The level of noise directly affects the performance of the data fusion algorithm. During the parameter debugging stage, understanding and considering the noise level will help optimize the robustness and accuracy of the algorithm.

[0057] Timestamp: For time series data, timestamp is a crucial feature. It records the time point when the data was collected, which helps to align the time series of different data sources during the data fusion process. The accuracy and consistency of timestamps are crucial to ensure the correctness and reliability of the fusion results. When debugging parameters, you may need to adjust parameters such as the time alignment algorithm and the size of the fusion window.

[0058] Spatial location: In spatially distributed data fusion, spatial location is a core feature. It describes the geographical location or spatial distribution of data points and is crucial for understanding the spatial relationship between data. During the parameter debugging stage, spatial location information can be used to optimize the spatial weight allocation and spatial interpolation methods in the fusion algorithm.

[0059] Data quality indicators: Data quality indicators such as completeness, accuracy, consistency, and timeliness are also important factors affecting the parameter debugging of data fusion algorithms. These indicators reflect the overall quality and availability of data and are essential to ensure the accuracy and reliability of fusion results. When debugging parameters, it may be necessary to adjust the parameter settings of the fusion algorithm according to the data quality indicators to optimize the fusion performance.

[0060] Correlation analysis: Correlation analysis is an important means to evaluate the correlation between different data sources. By calculating indicators such as correlation coefficients or mutual information between different features, we can understand the degree of association between them. In the parameter debugging stage, correlation analysis helps to identify key features and optimize the weight allocation and fusion strategy in the fusion algorithm.

[0061] Data distribution characteristics: Data distribution characteristics such as mean, variance, skewness and kurtosis are also important factors affecting the parameter debugging of data fusion algorithm. These characteristics reflect the statistical laws and distribution patterns of data, which are crucial for optimizing parameter settings and fusion strategies in fusion algorithms.

[0062] In one embodiment, the debugging parameters of the data fusion algorithm are determined according to the detected samples, and can be determined based on existing methods, such as establishing a debugging analysis model and performing intelligent analysis through the debugging analysis model.

[0063] Exemplarily, sample detection data from different dimensions and different detectors are collected; these data are preprocessed, including data cleaning, data format unification, data normalization, etc., to ensure data quality and consistency.

[0064] Useful features are extracted from the preprocessed data, which should be able to reflect the core information of the sample test data. Feature selection methods, such as statistics-based methods and machine learning-based methods, are used to screen out features that have an important impact on model building.

[0065] According to the complexity of the problem and the characteristics of the data, select the appropriate intelligent algorithm as the basis for debugging the analysis model, such as neural network, support vector machine, decision tree, etc. Use the training data set to train the model, adjust the parameters and structure of the model, so that it can accurately classify, predict or cluster the sample detection data.

[0066] Use the validation data set to validate the trained model and evaluate the model's performance, such as accuracy, recall, F1 score, etc. Optimize the model based on the validation results, including adjusting model parameters, adding features, and improving algorithms.

[0067] The optimized model is deployed to actual application scenarios for processing and analysis of sample detection data.

[0068] In one embodiment, when the debugging result cannot meet the user's requirements, the user can manually adjust it. However, in some cases, the user lacks the corresponding debugging experience and professional knowledge, resulting in the debugging result not meeting the requirements. Therefore, in this embodiment, remote debugging can also be performed, that is, when the debugging result of debugging the data fusion algorithm according to the debugging parameters does not meet the user's requirements, remote control debugging is performed; the remote control debugging method includes: When the user has the need for remote control, the data management permission is activated, that is, the platform is allowed to directly access and debug the data fusion algorithm; remote control information is sent to the platform, and the platform collects debugging and analysis data in real time after the data management permission is activated, that is, sample type, data fusion algorithm parameters, presentation effect and other related data; the debugging and analysis data is analyzed, the debugging parameters are determined, and the data fusion algorithm is debugged according to the debugging parameters.

[0069] In one embodiment, the debugging analysis data may be analyzed by the platform staff to determine the debugging parameters.

[0070] In one embodiment, the debugging analysis data is analyzed, and a debugging analysis model can be established based on a neural network or the like, and intelligent analysis can be performed through the debugging analysis model; that is, the debugging analysis model in the above embodiment.

[0071] For example, the platform establishes a remote connection with the user, and remote access tools such as VPN, remote desktop, SSH, etc. are used to ensure the security and stability of the connection for subsequent debugging work.

[0072] After the data management authority is activated, the sample test data at the user's site is transmitted to the platform's server in real time, or the platform's debugging tools and data sets are transmitted to the user's device. Ensure the integrity and accuracy of the data and avoid data loss or damage during the transmission process.

[0073] The platform uses debugging tools to remotely debug the user's equipment, including modifying codes, adjusting parameters, viewing logs, etc. The platform monitors data changes and system status during the debugging process in real time to identify and solve problems in a timely manner.

[0074] According to the debugging results and user feedback, the model is iteratively optimized to improve the performance and accuracy of the model. The optimized model is redeployed to the user's device for a new round of testing and verification. The platform can provide users with training and guidance to help them better understand and use data fusion algorithms and debug analysis models. Provide technical support and consulting services to solve problems and difficulties encountered by users during use.

[0075] Step 4: Fusing the sample test data through the debugged data fusion algorithm to obtain fused test data; Step 5: Obtain the user's sample analysis target, process the fusion detection data according to the sample analysis target, and obtain the sample analysis result.

[0076] For example, an analysis report can be generated based on the user's sample analysis objectives, including information such as sample composition and concentration. The analysis can be performed using existing data summarization and test report generation technologies.

[0077] A sample analysis system based on a multi-dimensional gas chromatograph data fusion algorithm, comprising an equipment analysis module, a detection module, an algorithm debugging module and a data analysis module; The equipment analysis module is used to analyze the multi-dimensional gas chromatograph and set the data fusion algorithm for application; The detection module is used to detect the detection sample and obtain corresponding sample detection data; The algorithm debugging module is used to debug the data fusion algorithm; The data analysis module is used to perform detection analysis, fuse the sample detection data through the debugged data fusion algorithm to obtain fused detection data; obtain the user's sample analysis target, process the fused detection data according to the sample analysis target, and obtain the sample analysis result.

[0078] The above formulas are all calculated by removing dimensions and taking numerical values. The formula is a formula that is closest to the actual situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained by simulating a large amount of data.

[0079] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A sample analysis method based on a multidimensional gas chromatograph data fusion algorithm, characterized in that: Methods include: Step 1: Collect multi-dimensional gas chromatograph information, and set a data fusion algorithm according to the multi-dimensional gas chromatograph information; Step 2: Identify the test sample, test the test sample by a multi-dimensional gas chromatograph, and obtain corresponding sample test data; Step 3: determining the debugging parameters of the data fusion algorithm according to the sample detection data, and debugging the data fusion algorithm according to the debugging parameters; Step 4: Perform fusion processing on the sample detection data through the debugged data fusion algorithm to obtain fusion detection data; Step 5: Obtain the user's sample analysis target, process the fusion detection data according to the sample analysis target, and obtain the sample analysis result.

2. A sample analysis method based on multidimensional gas chromatograph data fusion algorithm according to claim 1, characterized in that: The setting methods of data fusion algorithm include: Determining a sample detection range of the multidimensional gas chromatograph according to the multidimensional gas chromatograph information; Determine the potential fusion data according to the multi-dimensional gas chromatograph information and the sample detection range; perform a search according to the potential fusion data to obtain a plurality of fusion algorithms to be selected; The candidate fusion algorithms are screened according to the sample detection range to determine the data fusion algorithm to be applied.

3. A sample analysis method based on multidimensional gas chromatograph data fusion algorithm according to claim 2, characterized in that: Methods for screening the fusion algorithms to be selected according to the sample detection range include: Identify the sample types corresponding to the sample detection range, and count the detection proportions of the sample types; Performing a detection simulation on the sample type by using the candidate fusion algorithm to obtain a fusion effect value of data fusion performed by the candidate fusion algorithm on the sample type; The sample type is marked as i, i=1, 2, ..., n, and n is the number of sample types; The obtained fusion effect value is marked as RH i ; Mark the detection specific gravity of the sample type as δ i ; The screening value of the corresponding candidate fusion algorithm is calculated according to the screening formula. The screening formula is: ； Where: SW is the screening value; The candidate fusion algorithm with the largest screening value is selected as the applied data fusion algorithm.

4. The sample analysis method based on multidimensional gas chromatograph data fusion algorithm according to claim 3, characterized in that: Methods for simulating detection of sample types by using a candidate fusion algorithm include: Acquire fusion materials corresponding to the sample types, perform fusion processing on the fusion materials through the selected fusion algorithm, and obtain fusion result data; Preset fusion effect evaluation indicators and indicator sub-areas, and set corresponding reference standards and standard discount methods according to the fusion effect evaluation indicators; According to the fusion effect evaluation index, the fusion result data is collected to obtain the index analysis data corresponding to the fusion effect evaluation index; according to the reference standard and the standard discount method, the index analysis data is analyzed to obtain the index score corresponding to the fusion effect evaluation index; The index scores corresponding to each fusion effect evaluation index are accumulated to obtain the fusion effect value.

5. The sample analysis method based on multidimensional gas chromatograph data fusion algorithm according to claim 1, characterized in that: Methods for determining debugging parameters of the data fusion algorithm include: The platform sets a parameter representative curve, wherein the horizontal axis of the parameter representative curve is the serial number value and the vertical axis is the parameter representative value; each serial number value corresponds to a sample feature set data; each parameter representative value corresponds to a debugging reference parameter; Identify the sample detection data of the detection sample, identify the sample feature set data corresponding to the sample detection data, match the corresponding debugging reference parameters for the sample feature set data according to the parameter representative curve, and mark them as debugging parameters.

6. The sample analysis method based on multidimensional gas chromatograph data fusion algorithm according to claim 5, characterized in that: The parameter representative curve setting methods include: Determine the sample types that can be detected by the multidimensional gas chromatograph, obtain historical detection data of the sample types, determine historical sample detection data of the sample types according to the historical detection data, and mark them as sample material data; and determine a historical debugging parameter set corresponding to the sample material data according to the historical detection data; The platform presets each sample feature, performs feature processing on the sample material data according to the sample feature, and obtains sample feature set data corresponding to the sample material data; Identify the historical debugging parameter set corresponding to the sample feature set data, evaluate the fusion effect value corresponding to each historical debugging parameter in the historical debugging parameter set; generate a fusion effect curve according to the fusion effect value and the corresponding historical debugging parameter; determine the debugging reference parameter of the sample feature set data according to the fusion effect curve; Sorting the sample feature set data corresponding to the sample types to obtain a first sequence; setting corresponding debugging reference parameters for the corresponding sample feature set data in the first sequence; Setting corresponding sequence number values ​​for each sample feature set data according to the first sequence; Mark the debugging reference parameter with a serial number value of 1 as a benchmark parameter, and set corresponding parameter representative values ​​for each debugging reference parameter according to the benchmark parameter; A parameter representative curve is generated according to the serial number value and the parameter representative value, with the horizontal axis being the serial number value and the vertical axis being the parameter representative value.

7. The sample analysis method based on multidimensional gas chromatograph data fusion algorithm according to claim 6, characterized in that: The first sequence of settings includes: Step SA1: Determine the first ranked sample feature set data in a preset manner to form an initial sequence, and determine the benchmark data based on the initial sequence, where the benchmark data is the last ranked sample feature set data in the initial sequence; Step SA2: Calculate the similarity between the benchmark data and the remaining sample feature set data, add the sample feature set data with the highest similarity to the initial sequence, obtain a new initial sequence, and determine new benchmark data based on the new initial sequence; Step SA3: Loop step SA2 until there is no remaining sample feature set data, and mark the initial sequence as the first sequence.

8. The sample analysis method based on multidimensional gas chromatograph data fusion algorithm according to claim 1, characterized in that: When the debugging result of the data fusion algorithm according to the debugging parameters does not meet the user's requirements, the platform performs remote control debugging.

9. The sample analysis method based on multidimensional gas chromatograph data fusion algorithm according to claim 8, characterized in that: The methods for remote control debugging by the platform include: When the user has a need for remote control, the user opens the data management authority and sends remote control information to the platform; when the platform receives the remote control information, it collects debugging and analysis data in real time after the data management authority is opened; The debugging analysis data is analyzed to determine debugging parameters, and the data fusion algorithm is debugged according to the debugging parameters.

10. A sample analysis system based on a multidimensional gas chromatograph data fusion algorithm, characterized in that: A sample analysis method based on a multidimensional gas chromatograph data fusion algorithm according to any one of claims 1 to 9 is implemented, comprising a device analysis module, a detection module, an algorithm debugging module and a data analysis module; The equipment analysis module is used to analyze the multi-dimensional gas chromatograph and set the data fusion algorithm for application; The detection module is used to detect the detection sample and obtain corresponding sample detection data; The algorithm debugging module is used to debug the data fusion algorithm; The data analysis module is used to perform detection analysis, and fuse the sample detection data through the debugged data fusion algorithm to obtain fused detection data; Obtain the user's sample analysis target, process the fusion detection data according to the sample analysis target, and obtain the sample analysis result.

Images