Remote spectrophotometric detection method and system based on cloud platform

Abstract

The invention provides a remote spectrophotometric detection method and system based on a cloud platform. The method comprises the following steps: synchronously receiving spectral data from a plurality of remote spectrophotometers to obtain consistent sample spectral information; based on the consistent sample spectrum information, a calculation process is accelerated through the distributed calculation capability of a cloud platform, and a detailed component analysis report is generated; according to the detailed component analysis report, personalized guidance and suggestions are provided according to specific requirements of a user, and customized detection result notices and personalized guidance suggestions are generated; and dynamically adjusting an updating strategy of the spectrum database, discovering potential sample characteristics and detection modes through a data mining technology, and generating an optimized spectrum database and an optimized detection mode. According to the technical scheme provided by the invention, the spectrophotometric detection effect is improved, and the reliability and accuracy of the overall photometric detection result are improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of remote spectrophotometric detection, and in particular to a remote spectrophotometric detection method and system based on a cloud platform. Background Art

[0002] Remote spectrophotometric detection technology is widely used in modern medical care, environmental monitoring, food testing and other fields. These fields have strict requirements for accurate analysis of sample components, especially in shelter medical facilities, where rapid and accurate detection of multiple components in biological samples is essential for the diagnosis and treatment of diseases. In addition, with the development of the Internet of Things and cloud computing technologies, the demand for remote monitoring and data analysis is growing, requiring the detection system to be able to receive and process data from multiple remote spectrophotometers in real time and provide timely and accurate analysis results.

[0003] Currently, existing remote spectrophotometric detection methods mainly rely on a single or a few fixed-position spectrophotometers for data collection, and then perform component analysis through traditional spectral analysis algorithms. However, these systems usually lack correction processing for measurement differences between different instruments, resulting in poor data consistency and poor spectrophotometric detection results. Summary of the invention

[0004] The embodiments of the present application provide a remote spectrophotometric detection method and system based on a cloud platform, which are used to solve the problem of poor spectrophotometric detection effect in the prior art.

[0005] In the first aspect, the present application provides a remote spectrophotometric detection method based on a cloud platform, comprising:

[0006] Synchronously receive spectral data from multiple remote spectrophotometers, correct the measurement differences between different instruments, and obtain consistent sample spectral information;

[0007] Based on the consistent sample spectral information, a high-order spectral analysis algorithm combined with chemometrics technology is used to perform a detailed analysis of the concentration ratios of multiple components in the sample. The distributed computing capabilities of the cloud platform are used to accelerate the calculation process, improve the detection accuracy and speed, and generate a detailed component analysis report.

[0008] According to the detailed component analysis report, key information in the component analysis report is converted into an easy-to-understand text format, personalized guidance and suggestions are provided according to the specific needs of the user, and customized test result notifications and personalized guidance suggestions are generated;

[0009] Utilizing the customized test result notification and personalized guidance suggestions, combined with the machine learning algorithm, the update strategy of the spectral database is dynamically adjusted, the spectral analysis algorithm is continuously optimized, and potential sample characteristics and detection modes are discovered through data mining technology to generate an optimized spectral database and detection mode.

[0010] Optionally, based on the consistent sample spectral information, a high-order spectral analysis algorithm combined with chemometrics technology is used to perform a detailed analysis of the concentration ratios of multiple components in the sample, and the distributed computing capabilities of the cloud platform are used to accelerate the calculation process, improve the detection accuracy and speed, and generate a detailed component analysis report, including:

[0011] According to the consistent sample spectral information, a high-order spectral analysis algorithm is used to perform multi-dimensional analysis on the spectral data, extract characteristic spectral signals of various components in the sample, and generate a characteristic spectral signal set;

[0012] Based on the characteristic spectral signal set, a multivariate regression model is established in combination with chemometrics technology, and the concentration ratios of multiple components in the sample are calculated by the model to generate preliminary component analysis results;

[0013] Using the distributed computing capability of the cloud platform, the preliminary component analysis results are distributed to multiple computing nodes for parallel processing to generate optimized component analysis results;

[0014] According to the optimized component analysis results, combined with the data visualization tool of the cloud platform, a detailed component analysis report is generated.

[0015] Optionally, the method of performing multi-dimensional analysis on the spectral data based on the consistent sample spectral information using a high-order spectral analysis algorithm to extract characteristic spectral signals of various components in the sample and generate a characteristic spectral signal set includes:

[0016] In calculating the characteristic spectral signal intensity F i Previously, spectral data needed to be smoothed, baseline corrected, and normalized to ensure data accuracy and consistency;

[0017]

[0018] Among them, F i is the characteristic spectral signal intensity, corresponding to the characteristic spectral signal of the i-th component; I i is the original spectral intensity of the i-th component in the consistent sample spectral information; α is the amplitude factor of the Gaussian function; μ is the mean of the spectral intensity; σ is the standard deviation of the spectral intensity; β is the coefficient for adjusting the nonlinear term; γ is the nonlinear exponent; η is the amplitude factor of the sine function; λ is the wavelength of the sine function; θ is the amplitude factor of the exponential decay term; φ is the decay rate of the exponential decay term;

[0019] After calculating the characteristic spectrum signal intensity F i After that, the characteristic spectral signal is normalized and assigned a weight w i , calculate and generate the comprehensive evaluation score S j , reflecting the overall quality of the characteristic spectral signals of each component in the sample;

[0020]

[0021] Among them, S j is a comprehensive evaluation score used to evaluate the overall quality of the characteristic spectral signal set of the jth sample; n is the number of components in the sample; w i is the weight of the i-th component, reflecting the importance of the component in the sample; F i is the characteristic spectrum signal intensity of the i-th component; max(F) is the maximum value of all characteristic spectrum signal intensities; δ is the power of the normalized characteristic spectrum signal intensity; F avg is the average value of the characteristic spectrum signal intensity; τ is the variance of the characteristic spectrum signal intensity; K is the coefficient of the logarithmic function; min(F) is the minimum value of all characteristic spectrum signal intensities;

[0022] After calculating the comprehensive evaluation score S j After that, according to the comprehensive evaluation score S j The characteristic spectral signals are sorted and screened, cluster analysis is performed, characteristic spectral signal clusters are generated, and they are organized into structured data sets.

[0023] Optionally, the method of establishing a multivariate regression model based on the characteristic spectral signal set in combination with chemometrics technology, and calculating the concentration ratios of multiple components in the sample through the model to generate preliminary component analysis results, including:

[0024] According to the characteristic spectral signal set, a suitable chemometric method is selected to preprocess the characteristic spectral signal, remove noise and irrelevant variables, and obtain an optimized characteristic spectral signal;

[0025] Based on the optimized characteristic spectral signal and in combination with known standard sample data, a multivariate regression model capable of establishing a mathematical relationship between the characteristic spectral signal and the concentration ratios of multiple components in the sample is established to generate a model training result;

[0026] Using the model training results, the characteristic spectral signals of unknown samples are predicted, and the concentration ratios of multiple components in the samples are calculated through a multivariate regression model to generate prediction results;

[0027] Based on the prediction results, combined with the confidence interval and error analysis of the model, preliminary component analysis results are generated.

[0028] Optionally, the optimized component analysis results are combined with a data visualization tool on a cloud platform to generate a detailed component analysis report, including:

[0029] According to the optimized component analysis results, the data processing function of the cloud platform is used to further count and summarize the analysis results, calculate the average concentration, standard deviation and confidence interval of each component, and generate statistical analysis data;

[0030] Based on the statistical analysis data, combined with the data visualization tool of the cloud platform, an intuitive visualization chart is generated;

[0031] Generate detailed report text using the visual chart in combination with detailed text description;

[0032] Based on the detailed report text and the visual chart, combined with the user's historical test records and preference settings, the cloud platform's intelligent recommendation system provides the user with a customized report interpretation service and generates a detailed component analysis report.

[0033] Optionally, the customized detection result notification and personalized guidance suggestions are used in combination with a machine learning algorithm to dynamically adjust the update strategy of the spectral database, continuously optimize the spectral analysis algorithm, discover potential sample characteristics and detection modes through data mining technology, and generate an optimized spectral database and detection mode, including:

[0034] According to the customized test result notification and personalized guidance suggestions, user feedback information is collected to generate a user feedback data set;

[0035] Based on the user feedback data set, the user feedback information is analyzed in combination with a machine learning algorithm to identify common problems and improvement directions, dynamically adjust the update strategy of the spectral database to ensure the timeliness and accuracy of the database content, and generate an update strategy adjustment plan;

[0036] Using the update strategy adjustment scheme, regularly updating the spectral database, adding new standard spectral data and sample information, and removing obsolete and invalid data, to generate an optimized spectral database;

[0037] Based on the optimized spectral database, a large amount of spectral data is analyzed through data mining technology to discover potential sample characteristics and detection modes. Combined with the latest research results and industry standards, the spectral analysis algorithm is continuously optimized to generate optimized detection modes.

[0038] Optionally, based on the user feedback data set, the user feedback information is analyzed in combination with a machine learning algorithm, common problems and improvement directions are identified, the update strategy of the spectral database is dynamically adjusted to ensure the timeliness and accuracy of the database content, and an update strategy adjustment plan is generated, including:

[0039] When calculating the recognition probability P of the kth common problem k Previously, the user feedback data was cleaned, sentiment analyzed, and keywords extracted to generate a relevance score r i and severity scores i , and count the frequency of problems f ik And record submission time t i ;

[0040]

[0041] Among them, P k is the recognition probability of the kth common problem; m is the total number of user feedback information; ri is the relevance score of the i-th user feedback information; f ik is the frequency of occurrence of the kth type of problem in the i-th user feedback information; λ is the weight factor of the logarithmic term, which is used to adjust the impact of the logarithmic term; s i is the severity score of the i-th user feedback information; η is the amplitude factor of the time impact item, which is used to adjust the impact of the time impact item; t i is the submission time of the i-th user feedback information, which is used to consider the time factor;

[0042] After calculating the recognition probability P of the kth common problem k Then, calculate the average P of all common problem recognition probabilities avg and standard deviation σ P , and combined with the time factor t to calculate the update strategy adjustment score U t :

[0043]

[0044] Among them, U t is the update strategy adjustment score, which is used to decide whether the spectral database needs to be updated; t is the time since the last update; α is the amplitude factor of the Gaussian function, which is used to adjust the height of the Gaussian peak; P k is the recognition probability of the kth common problem; P avg is the average of all common problem recognition probabilities; σ P is the standard deviation of the probability of identifying common problems; β is the amplitude factor of the time decay term, which is used to adjust the influence of time; τ is the decay rate of the time decay term, which reflects the speed of the influence of time; γ is the coefficient of the weighted average term, which is used to adjust the influence of the weighted average term; n is the number of categories of common problems; w k is the weight of the kth type of common problem, reflecting the importance and urgency of the problem;

[0045] Calculate U t After that, set the threshold value, when Ut When the threshold is exceeded, the update process of the spectral database is triggered. k and the weight of the common problem w k , identify common problem categories that need to be updated first and develop specific updates.

[0046] Optionally, the update strategy adjustment scheme is used to regularly update the spectral database, add new standard spectral data and sample information, and remove obsolete and invalid data to generate an optimized spectral database, including:

[0047] According to the update strategy adjustment plan, set the time interval and trigger conditions for regular updates and generate an update plan;

[0048] Based on the update plan, start an update process of the spectral database, collect the latest standard spectral data and sample information from multiple data sources, and generate a data set to be updated;

[0049] Using the data set to be updated, quality control is performed on the collected new data, and data cleaning and verification algorithms are used to remove noise and erroneous data to generate a high-quality new data set;

[0050] Based on the high-quality new data set, combined with the existing spectral database, data fusion technology is used to integrate the new data with the existing data, and the data validity evaluation algorithm is used to identify and remove outdated and invalid data to generate an optimized spectral database.

[0051] Optionally, according to the optimized spectral database, a large amount of spectral data is analyzed by data mining technology to discover potential sample characteristics and detection modes, and the spectral analysis algorithm is continuously optimized in combination with the latest research results and industry standards to generate an optimized detection mode, including:

[0052] Based on the optimized spectral database, a large amount of spectral data is deeply analyzed using data mining technology to identify potential characteristic patterns and abnormal phenomena in the sample and generate potential characteristic analysis results;

[0053] Based on the potential characteristic analysis results, combined with the latest research results and industry standards, the existing spectrum analysis algorithm is evaluated and improved, and new algorithm models and technical means are introduced to generate an improved spectrum analysis algorithm;

[0054] Using the improved spectral analysis algorithm, reanalyze the sample data in the optimized spectral database, verify the effect of the algorithm improvement, and generate an algorithm performance evaluation report;

[0055] According to the algorithm performance evaluation report, combined with user feedback and actual detection results, the spectral analysis algorithm is continuously optimized to form an optimized detection mode.

[0056] In a second aspect, the present application provides a remote spectrophotometric detection system based on a cloud platform, comprising:

[0057] A receiving module is used to synchronously receive spectral data from multiple remote spectrophotometers, correct the measurement differences between different instruments, and obtain consistent sample spectral information;

[0058] An analysis module is used to perform a detailed analysis of the concentration ratios of multiple components in the sample based on the consistent sample spectral information using a high-order spectral analysis algorithm combined with chemometrics technology, accelerate the calculation process through the distributed computing capability of the cloud platform, improve the detection accuracy and speed, and generate a detailed component analysis report;

[0059] A conversion module, for converting key information in the component analysis report into an easy-to-understand text format according to the detailed component analysis report, providing personalized guidance and suggestions according to the specific needs of the user, and generating customized test result notifications and personalized guidance suggestions;

[0060] The adjustment module is used to utilize the customized detection result notification and personalized guidance suggestions, combine with the machine learning algorithm to dynamically adjust the update strategy of the spectral database, continuously optimize the spectral analysis algorithm, discover potential sample characteristics and detection modes through data mining technology, and generate an optimized spectral database and detection mode.

[0061] In a third aspect, the present application provides a computing device comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a remote spectrophotometric detection method based on a cloud platform as described in the first aspect.

[0062] In a fourth aspect, the present application provides a computer storage medium storing a computer program, which, when executed by a computer, implements a remote spectrophotometric detection method based on a cloud platform as described in the first aspect.

[0063] In an embodiment of the present application, spectral data are synchronously received from multiple remote spectrophotometers, and measurement differences between different instruments are corrected to obtain consistent sample spectral information; based on the consistent sample spectral information, a high-order spectral analysis algorithm is used in combination with chemometrics technology to perform a detailed analysis of the concentration ratios of multiple components in the sample, and the distributed computing capabilities of the cloud platform are used to accelerate the calculation process, improve detection accuracy and speed, and generate a detailed component analysis report; based on the detailed component analysis report, the key information in the component analysis report is converted into an easy-to-understand text format, and personalized guidance and suggestions are provided according to the specific needs of the user, and customized detection result notifications and personalized guidance suggestions are generated; using the customized detection result notifications and personalized guidance suggestions, combined with a machine learning algorithm, the update strategy of the spectral database is dynamically adjusted, the spectral analysis algorithm is continuously optimized, and potential sample characteristics and detection modes are discovered through data mining technology, and an optimized spectral database and detection mode are generated.

[0064] The technical solution of this application has the following beneficial effects:

[0065] By synchronously receiving spectral data from multiple remote spectrophotometers and correcting the measurement differences between different instruments, the consistency and accuracy of the sample spectral information is ensured. This helps to reduce data deviations caused by instrument differences and improve the reliability of the overall test results; the use of high-order spectral analysis algorithms combined with chemometrics technology can perform detailed analysis of the concentration ratios of multiple components in the sample. The distributed computing power of the cloud platform accelerates the calculation process, improves the detection accuracy and speed, and generates a detailed component analysis report. This makes the detection process more efficient and can quickly provide accurate results; based on the detailed component analysis report, key information is converted into an easy-to-understand text format, and personalized guidance and suggestions are provided according to the user's specific needs, generating customized test result notifications and personalized guidance suggestions. This user-friendly approach enhances user understanding and satisfaction, enabling users to better utilize test results; using customized test result notifications and personalized guidance suggestions, combined with machine learning algorithms to dynamically adjust the update strategy of the spectral database, and continuously optimize the spectral analysis algorithm. This ensures the timeliness and accuracy of the database content. At the same time, potential sample characteristics and detection modes are discovered through data mining technology, further improving the system's intelligence level and detection capabilities. By continuously optimizing the spectral analysis algorithm and generating optimized detection modes, the system can better adapt to different application scenarios and detection needs. This not only improves the flexibility of the system, but also enhances its adaptability in complex environments, ensuring long-term service quality and accuracy.

[0066] Based on the consistent sample spectral information, a high-order spectral analysis algorithm is used to perform multi-dimensional analysis of the spectral data and extract the characteristic spectral signal set; a multivariate regression model is established in combination with chemometrics technology to calculate the concentration ratios of multiple components in the sample and generate preliminary component analysis results; the preliminary results are processed in parallel through the distributed computing capabilities of the cloud platform to generate optimized component analysis results; finally, a detailed component analysis report is generated in combination with the data visualization tools of the cloud platform.

[0067] Furthermore, through high-order spectral analysis algorithms and chemometrics technology, multiple components in the sample can be extracted and analyzed more accurately, improving the accuracy of the test results; utilizing the distributed computing capabilities of the cloud platform, the preliminary analysis results can be processed in parallel, significantly improving the calculation speed and shortening the test time; combining data visualization tools to generate detailed component analysis reports makes the test results more intuitive and easy to understand, making it easier for users to understand and use; the generated detailed component analysis reports not only contain rich data, but also provide clear charts and explanations, enhancing user understanding and satisfaction.

[0068] Based on customized test result notifications and personalized guidance suggestions, user feedback information is collected to generate a user feedback data set; user feedback is analyzed in conjunction with machine learning algorithms to identify common problems and improvement directions, dynamically adjust the update strategy of the spectral database, and generate an update strategy adjustment plan; the spectral database is updated regularly, new standard spectral data and sample information are added, outdated and invalid data are removed, and an optimized spectral database is generated; a large amount of spectral data is analyzed through data mining technology to discover potential sample characteristics and detection modes, and the spectral analysis algorithm is continuously optimized in combination with the latest research results and industry standards to generate optimized detection modes.

[0069] Furthermore, by dynamically adjusting the update strategy, the timeliness and accuracy of the spectral database content are ensured to promptly reflect the latest research results and industry standards; by utilizing user feedback information, customized test result notifications and personalized guidance suggestions are provided, thereby improving user satisfaction and the practicality of the system; through data mining technology, potential sample characteristics and detection patterns are discovered, and combined with the latest research results, the spectral analysis algorithm is continuously optimized to improve detection accuracy and efficiency; by combining machine learning algorithms and data mining technology, the system can automatically identify problems and perform self-optimization, thereby enhancing the intelligence and adaptability of the system.

[0070] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0072] Figure 1 A flowchart of a remote spectrophotometric detection method based on a cloud platform provided in an embodiment of the present application;

[0073] Figure 2 A schematic diagram of the structure of a remote spectrophotometric detection system based on a cloud platform provided in an embodiment of the present application;

[0074] Figure 3 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0075] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0076] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0077] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0078] Figure 1 A flowchart of a remote spectrophotometric detection method based on a cloud platform is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0079] 101. Synchronously receive spectral data from multiple remote spectrophotometers, correct the measurement differences between different instruments, and obtain consistent sample spectral information;

[0080] Synchronous reception of spectral data refers to the real-time acquisition of spectral data from multiple remote spectrophotometers distributed in the network. These data include the absorption or reflection intensity of the sample at different wavelengths, which is used to analyze the composition and characteristics of the sample. Measurement difference correction processing refers to the elimination of systematic errors and deviations between different instruments through algorithms to ensure the consistency and comparability of all data.

[0081] This step first synchronously receives spectral data from multiple remote spectrophotometers, and then uses correction algorithms (such as cross-validation, standardization, etc.) to correct the measurement differences between different instruments, and finally generates consistent sample spectral information.

[0082] Assume that a shelter medical facility is equipped with three spectrophotometers of different models (A, B, C), and each instrument measures the spectral data of the same sample. Due to slight differences between the instruments, the measurement results may not be completely consistent. To eliminate these differences, the following steps can be taken:

[0083] First, a set of standard samples with known components are used to measure on three instruments respectively, and the spectral data of each instrument is recorded; then, the measurement data of the standard samples are cross-validated and the correction coefficient of each instrument is calculated. For example, through a linear regression model, the relationship between each instrument and the standard value is found; further, the calculated correction coefficient is used to correct the measurement data of the actual sample; finally, the corrected data is checked for consistency again to ensure that the data of all instruments are at the same level. If consistency is still a problem, the correction coefficient can be further adjusted or a more complex correction method can be used, such as polynomial fitting or nonlinear correction.

[0084] Through the above steps, the systematic errors and deviations between different instruments can be effectively eliminated, ensuring the consistency and comparability of all data, thereby obtaining consistent sample spectral information.

[0085] 102. Based on the consistent sample spectral information, a high-order spectral analysis algorithm combined with chemometrics technology is used to perform a detailed analysis of the concentration ratios of multiple components in the sample, and the distributed computing capabilities of the cloud platform are used to accelerate the calculation process, improve the detection accuracy and speed, and generate a detailed component analysis report;

[0086] High-order spectral analysis algorithm is an advanced data processing method that can extract characteristic information of various components in a sample from complex spectral signals. Chemometrics technology is a set of statistical and mathematical methods used to establish a relationship model between sample components and spectral signals. The distributed computing capability of the cloud platform refers to the use of cloud computing resources to process large amounts of data in parallel and speed up computing.

[0087] This step is based on consistent sample spectral information, uses high-order spectral analysis algorithms to extract characteristic spectral signals, and combines chemometrics technology to establish a multivariate regression model to calculate the concentration ratios of multiple components in the sample. The distributed computing capabilities of the cloud platform accelerate the calculation process and ultimately generate a detailed component analysis report.

[0088] Optionally, in step 102, based on the sample spectral information of the consistency, a high-order spectral analysis algorithm combined with chemometrics technology is used to perform a detailed analysis of the concentration ratios of multiple components in the sample, and the distributed computing capability of the cloud platform is used to accelerate the computing process, improve the detection accuracy and speed, and generate a detailed component analysis report, including:

[0089] According to the consistent sample spectral information, a high-order spectral analysis algorithm is used to perform a multi-dimensional analysis of the spectral data, and the characteristic spectral signals of various components in the sample are extracted to generate a characteristic spectral signal set; based on the characteristic spectral signal set, a multivariate regression model is established in combination with chemometrics technology, and the concentration ratios of various components in the sample are calculated through the model to generate preliminary component analysis results; using the distributed computing capabilities of the cloud platform, the preliminary component analysis results are distributed to multiple computing nodes for parallel processing to generate optimized component analysis results; based on the optimized component analysis results, combined with the data visualization tools of the cloud platform, a detailed component analysis report is generated.

[0090] Wherein, the component analysis results of the optimization are combined with the data visualization tool of the cloud platform to generate a detailed component analysis report, including:

[0091] According to the optimized component analysis results, the data processing function of the cloud platform is used to further count and summarize the analysis results, calculate the average concentration, standard deviation and confidence interval of each component, and generate statistical analysis data; based on the statistical analysis data, combined with the data visualization tool of the cloud platform, generate intuitive visualization charts; use the visualization charts, combined with detailed text descriptions, to generate detailed report texts; based on the detailed report text and the visualization charts, combined with the user's historical detection records and preference settings, through the intelligent recommendation system of the cloud platform, provide users with customized report interpretation services and generate detailed component analysis reports.

[0092] Wherein, based on the characteristic spectral signal set, a multivariate regression model is established in combination with chemometrics technology, and the concentration ratios of multiple components in the sample are calculated by the model to generate preliminary component analysis results, including:

[0093] According to the characteristic spectral signal set, a suitable chemometric method is selected to preprocess the characteristic spectral signals, remove noise and irrelevant variables, and obtain optimized characteristic spectral signals; based on the optimized characteristic spectral signals, combined with known standard sample data, a multivariate regression model is established that can establish a mathematical relationship between the characteristic spectral signals and the concentration ratios of multiple components in the sample, and a model training result is generated; using the model training results, the characteristic spectral signals of unknown samples are predicted, and the concentration ratios of multiple components in the sample are calculated by the multivariate regression model to generate prediction results; based on the prediction results, combined with the confidence interval and error analysis of the model, a preliminary component analysis result is generated.

[0094] Suppose a shelter medical facility uses a remote spectrophotometric detection system to analyze the components of a patient's blood sample and generate a detailed component analysis report. The following is the specific implementation process:

[0095] First, according to the consistent sample spectral information, a high-order spectral analysis algorithm is used to perform a multi-dimensional analysis on the spectral data, and characteristic spectral signals of various components in the sample are extracted to generate a characteristic spectral signal set: the system receives the blood sample spectral data that is synchronously received and calibrated from multiple remote spectrophotometers; a high-order spectral analysis algorithm (such as principal component analysis PCA, independent component analysis IC A, etc.) is used to perform a multi-dimensional analysis on the spectral data, and characteristic spectral signals of various components (such as glucose, cholesterol, hemoglobin, etc.) in the blood sample are extracted; a characteristic spectral signal set is generated, which includes characteristic spectral signals of each component at different wavelengths;

[0096] Secondly, based on the characteristic spectral signal set, a multivariate regression model is established in combination with chemometric technology, and the concentration ratios of multiple components in the sample are calculated through the model to generate preliminary component analysis results: select appropriate chemometric methods (such as partial least squares PLS, principal component regression PCR, etc.), pre-process the characteristic spectral signals, remove noise and irrelevant variables, and obtain optimized characteristic spectral signals; combine known standard sample data (such as spectral data and component concentrations of standard blood samples) to establish a multivariate regression model that can establish a mathematical relationship between the characteristic spectral signals and the concentration ratios of multiple components in the sample; use the trained multivariate regression model to predict the characteristic spectral signals of unknown blood samples and calculate the concentration ratios of various components in the blood samples; based on the prediction results, combined with the confidence interval and error analysis of the model, generate preliminary component analysis results, including the concentration ratios of each component and their confidence intervals.

[0097] Furthermore, by utilizing the distributed computing capability of the cloud platform, the preliminary component analysis results are distributed to multiple computing nodes for parallel processing to generate optimized component analysis results: the preliminary component analysis results are distributed to multiple computing nodes on the cloud platform for parallel processing to accelerate the computing process; through parallel computing, the component analysis results are further optimized to improve detection accuracy and speed; and optimized component analysis results are generated, including more accurate component concentration ratios and confidence intervals.

[0098] Furthermore, based on the optimized component analysis results, combined with the data visualization tools of the cloud platform, a detailed component analysis report is generated: the optimized component analysis results are further statistically summarized and summarized using the data processing function of the cloud platform, the average concentration, standard deviation and confidence interval of each component are calculated, and statistical analysis data are generated; based on the statistical analysis data, combined with the data visualization tools of the cloud platform (such as chart generation tools), intuitive charts such as bar charts, line charts and scatter charts are generated to display the concentration distribution and change trend of each component; using the generated visualization charts, combined with detailed text descriptions, a detailed report text is written, including the concentration ratio of each component, statistical analysis results and chart explanations.

[0099] Finally, based on the detailed report text and visual charts, combined with the user's historical test records and preference settings, the cloud platform's intelligent recommendation system provides users with customized report interpretation services and generates detailed component analysis reports. For example, if the user has a history of diabetes, the system will emphasize changes in glucose concentration in the report and provide corresponding health advice.

[0100] Through the above steps, the system can efficiently and accurately generate detailed component analysis reports and provide personalized report interpretation services, thereby improving user understanding and satisfaction.

[0101] This application takes into account that in spectral analysis, the extraction and evaluation of characteristic spectral signals are key steps. Through high-order spectral analysis algorithms, characteristic spectral signals of various components in the sample can be extracted more accurately, and the overall quality of these signals can be reflected through comprehensive evaluation scores. This helps to improve the accuracy and reliability of subsequent component analysis.

[0102] Optionally, the method of performing multi-dimensional analysis on the spectral data based on the consistent sample spectral information using a high-order spectral analysis algorithm to extract characteristic spectral signals of various components in the sample and generate a characteristic spectral signal set includes:

[0103] Before calculating the characteristic spectral signal intensity Fi, the spectral data needs to be smoothed, baseline corrected, and normalized to ensure the accuracy and consistency of the data;

[0104]

[0105] Among them, F i is the characteristic spectral signal intensity, corresponding to the characteristic spectral signal of the i-th component; I i is the original spectral intensity of the i-th component in the consistent sample spectral information; α is the amplitude factor of the Gaussian function; μ is the mean of the spectral intensity; σ is the standard deviation of the spectral intensity; β is the coefficient for adjusting the nonlinear term; γ is the nonlinear exponent; η is the amplitude factor of the sine function; λ is the wavelength of the sine function; θ is the amplitude factor of the exponential decay term; φ is the decay rate of the exponential decay term;

[0106] After calculating the characteristic spectrum signal intensity F i After that, the characteristic spectral signal is normalized and assigned a weight w i , calculate and generate the comprehensive evaluation score S j , reflecting the overall quality of the characteristic spectral signals of each component in the sample;

[0107]

[0108] Among them, S j is a comprehensive evaluation score used to evaluate the overall quality of the characteristic spectral signal set of the jth sample; n is the number of components in the sample; w i is the weight of the i-th component, reflecting the importance of the component in the sample; F i is the characteristic spectrum signal intensity of the i-th component; max(F) is the maximum value of all characteristic spectrum signal intensities; δ is the power of the normalized characteristic spectrum signal intensity; F avg is the average value of the characteristic spectrum signal intensity; τ is the variance of the characteristic spectrum signal intensity; K is the coefficient of the logarithmic function; min(F) is the minimum value of all characteristic spectrum signal intensities;

[0109] After calculating the comprehensive evaluation score S j After that, according to the comprehensive evaluation score S j The characteristic spectral signals are sorted and screened, cluster analysis is performed, characteristic spectral signal clusters are generated, and they are organized into structured data sets.

[0110] This formula is designed to accurately extract and evaluate the characteristic spectral signals of each component in the sample during spectral analysis, thereby improving the accuracy and reliability of subsequent component analysis; through the combined use of these two formulas, a complete spectral signal processing and evaluation system is formed from feature extraction to comprehensive evaluation, making spectral data analysis more accurate and reliable, thereby better serving the needs of component analysis in practical applications.

[0111] The following is a brief introduction to the design reasons of each sub-item of the formula:

[0112]

[0113] Gaussian function term The Gaussian function is used to capture the peak characteristics in the spectral signal and is suitable for describing the main components in the spectrum. It can effectively identify and extract characteristic signals with significant peaks in the spectrum. By adjusting the amplitude factor α, mean μ and standard deviation σ, the main peak of the spectral signal can be accurately fitted, thereby improving the accuracy of feature extraction.

[0114] Nonlinear terms Nonlinear terms are used to capture the nonlinear characteristics of spectral signals and enhance the expressive power of the model. In actual spectral data, there may be a nonlinear relationship between the concentration of a component and the spectral intensity. By introducing nonlinear terms and adjusting the coefficient β and exponent γ, these nonlinear changes can be better fitted, improving the adaptability and accuracy of the model to complex spectral signals.

[0115] Sine function term The sine function is used to introduce periodic changes and capture the periodic features in the spectral signal. In some cases, the spectral signal may show periodic fluctuations or oscillations. By introducing the sine function and adjusting the amplitude factor η and wavelength λ, these periodic changes can be simulated, thereby more comprehensively describing the characteristics of the spectral signal.

[0116] Exponential decay term The exponential decay term is used to simulate the phenomenon that the spectral signal gradually saturates as the intensity increases. In practical applications, as the spectral intensity increases, the signal may gradually tend to saturation and no longer grow linearly. By introducing the exponential decay term and adjusting the amplitude factor θ and the decay rate φ, this saturation effect can be effectively simulated, making the model more consistent with the behavior of actual spectral data.

[0117] The following is a brief introduction to how to obtain the parameters of the formula:

[0118] α (amplitude factor of Gaussian function) optimizes the loss function by gradient descent method, and adjusts α to minimize the prediction error; μ (mean of spectral intensity) optimizes the loss function by gradient descent method, and adjusts μ to minimize the prediction error. The initial value can be set to the average value of spectral intensity in training data; σ (standard deviation of spectral intensity) optimizes the loss function by gradient descent method, and adjusts σ to minimize the prediction error. The initial value can be set to the standard deviation of spectral intensity in training data; β (coefficient of nonlinear term) optimizes the loss function by gradient descent method, and adjusts β to minimize the prediction error. The initial value can be set to 1.0; y (nonlinear exponent) optimizes the loss function by gradient descent method, and adjusts Adjust γ to minimize the prediction error, and the initial value can be set to 2.0; η (the amplitude factor of the sine function) optimizes the loss function through the gradient descent method, adjusts η to minimize the prediction error, and the initial value can be set to 1.0; λ (the wavelength of the sine function) optimizes the loss function through the gradient descent method, adjusts λ to minimize the prediction error, and the initial value can be set according to experience, such as 50; θ (the amplitude factor of the exponential decay term) optimizes the loss function through the gradient descent method, adjusts θ to minimize the prediction error, and the initial value can be set to 1.0; φ (the decay rate of the exponential decay term) optimizes the loss function through the gradient descent method, adjusts φ to minimize the prediction error, and the initial value can be set to 10.

[0119] The following is a brief introduction to the design reasons of each sub-item of the formula:

[0120]

[0121] Normalized characteristic spectral signal intensity By converting the characteristic spectral signal intensity F i Normalization to the maximum value max(F) ensures that the characteristic spectral signals of different components are compared on the same scale. The power δ can adjust the weight distribution of the normalized signal to make some important features more prominent.

[0122] Gaussian function term Gaussian function is used to capture the characteristic spectral signal intensity and the average value F avg By adjusting the variance τ, we can control the degree of penalty for deviation from the mean. This helps to highlight those signals that are close to the mean and representative.

[0123] Logarithmic term The logarithmic term is used to enhance the sensitivity to low-intensity signals and balance the signal strength distribution. The coefficient K controls the contribution of the logarithmic term, while min(F) ensures that the logarithmic term is still effective even at very low signal strengths.

[0124] Weight w i : Weight w iIt reflects the importance of the i-th component in the sample. By assigning different weights, the influence of certain components can be emphasized or weakened, so as to more reasonably evaluate the overall quality.

[0125] The following is a brief introduction to how to obtain the parameters of the formula:

[0126] w i (Weight) is set according to the importance of each component in the sample, and can be determined by expert knowledge, historical data, or using machine learning methods (such as principal component analysis PCA). δ (the power of the normalized characteristic spectral signal intensity) is optimized by gradient descent, and δ is adjusted to minimize the prediction error; the initial value can be set to 1.0 and adjusted according to the training data; F avg (the average value of the characteristic spectrum signal intensity) directly calculates the average value of all characteristic spectrum signal intensities from the training data; τ (the variance of the characteristic spectrum signal intensity) optimizes the loss function by the gradient descent method, and adjusts τ to minimize the prediction error; the initial value can be set to 1.0 and adjusted according to the training data; κ (the coefficient of the logarithmic function) optimizes the loss function by the gradient descent method, and adjusts K to minimize the prediction error; the initial value can be set to 0.1 and adjusted according to the training data; max(F) (the maximum value of all characteristic spectrum signal intensities) directly finds the maximum value of all characteristic spectrum signal intensities from the training data; min(F) (the minimum value of all characteristic spectrum signal intensities) directly finds the minimum value of all characteristic spectrum signal intensities from the training data; suppose there is a sample containing 3 components, and its spectral data is I1=100, I2=80, I3=120;

[0127] The parameters are set as Gaussian function parameters α = 1.5, μ = 100, σ = 10; nonlinear term parameters β = 0.5, γ = 2; sine function parameters η = 0.3, λ = 50; exponential decay term parameters θ = 0.2, φ = 5; weights w1 = 0.4, W2 = 0.3, w3 = 0.3; comprehensive evaluation parameters δ = 1.5, τ = 10, K = 0.1

[0128] Calculate the characteristic spectrum signal intensity F i :

[0129]

[0130] Normalization:

[0131] max(F)=7200.7027

[0132] min(F)=3200.1033

[0133]

[0134] Calculate the comprehensive evaluation score S j :

[0135]

[0136] Comprehensive evaluation score:

[0137] S j =S1+S2+S3=0.040818+0.0301+0.0544=0.125318

[0138] Suppose a threshold is set to determine the comprehensive evaluation score Sj. For example, suppose the threshold is 0.5: If Sj is greater than or equal to 0.5, this indicates that the overall quality of the characteristic spectral signals of the components in the sample is high, and the signal intensity and distribution are relatively ideal. If Sj is less than 0.5, this indicates that although the characteristic spectral signal intensity of some components is high, the overall quality may be affected by certain factors, resulting in a low comprehensive evaluation score; in this example, the comprehensive evaluation score Sj = 0.125318, which is far below the set threshold of 0.5.

[0139] Therefore, the following conclusions can be drawn: Since Sj is less than 0.5, it shows that although the characteristic spectral signal intensities of components 1 and 3 are high, the overall quality is affected by certain factors. These factors may include: there may be noise in the spectral data, which affects the quality of the signal; other components or external factors may interfere with the signal; there may be some factors that are not fully considered by the model, such as nonlinear effects and periodic changes in the spectral data.

[0140] Through this embodiment, we can see how the formula is applied to actual spectral data, and obtain the characteristic spectral signal intensity and comprehensive evaluation score through calculation, which provides a basis for further clustering analysis and generating structured data sets.

[0141] 103. Based on the detailed component analysis report, convert key information in the component analysis report into an easy-to-understand text format, provide personalized guidance and suggestions based on the specific needs of the user, and generate customized test result notifications and personalized guidance suggestions;

[0142] The component analysis report contains the concentration ratio of various components in the sample and the relevant analysis results. These professional terms and data are converted into easy-to-understand text formats to enable users to understand and use this information more intuitively. Personalized guidance and suggestions are customized services provided according to the specific needs of users to help users make better use of test results.

[0143] This step is based on a detailed component analysis report and uses natural language generation technology to convert key information into an easy-to-understand text format. It also combines a virtual expert system to provide personalized guidance and suggestions based on the user's specific needs, generating customized test result notifications and personalized guidance suggestions.

[0144] Suppose a shelter medical facility uses a remote spectrophotometric detection system to test a patient's blood sample and generate a detailed component analysis report. The following is the specific implementation process:

[0145] First, the system receives a detailed component analysis report from the cloud platform, which contains the concentration ratios of various components (such as glucose, cholesterol, hemoglobin, etc.) in the patient's blood sample and related analysis results.

[0146] Secondly, extract key information from the component analysis report, such as: glucose concentration: 95mg / dL (normal range: 70-99mg / dL); cholesterol concentration: 200mg / dL (normal range: <200mg / dL); hemoglobin concentration: 14g / dL (normal range: 13.8-17.2g / dL)

[0147] Then, natural language generation technology is used to convert the above key information into an easy-to-understand text format. For example: "Your blood test results show that the glucose concentration is 95mg / dL, which is within the normal range."; "Your cholesterol concentration is 200mg / dL, which is slightly higher than the normal range. It is recommended that you pay attention to your diet and have regular checkups."; "Your hemoglobin concentration is 14g / dL, which is within the normal range."

[0148] Furthermore, according to the specific needs of the user (for example, whether the patient has a history of diabetes, whether special dietary advice is needed, etc.), the virtual expert system is combined for analysis. For example, if the user has a history of diabetes, the system will pay special attention to the glucose concentration and provide more detailed advice.

[0149] Furthermore, personalized guidance suggestions are generated based on the user's specific situation. For example: "Given your history of diabetes, it is recommended that you continue to maintain a low-sugar diet and monitor your blood sugar levels regularly."; "Your cholesterol is slightly high, so it is recommended that you reduce your intake of saturated fat and cholesterol and increase your intake of dietary fiber, such as eating more vegetables and fruits."; "Your hemoglobin level is normal, but it is recommended that you maintain good living habits, such as moderate exercise and adequate sleep."

[0150] Finally, the above information is integrated into a customized test result notification, including key information and personalized guidance recommendations.

[0151] Through the above steps, the system can convert complex component analysis reports into easy-to-understand text format, provide personalized guidance and suggestions based on the user's specific needs, and generate customized test result notifications, thereby improving user understanding and satisfaction.

[0152] 104. Utilize the customized test result notification and personalized guidance suggestions, combine with the machine learning algorithm to dynamically adjust the update strategy of the spectral database, continuously optimize the spectral analysis algorithm, discover potential sample characteristics and detection modes through data mining technology, and generate an optimized spectral database and detection mode.

[0153] Customized test result notifications and personalized guidance recommendations contain user feedback information that can be used to improve system performance. Machine learning algorithms can automatically identify patterns and optimize themselves, and data mining techniques can discover hidden patterns and regularities from large amounts of data. Optimized spectral databases and detection modes refer to databases and detection methods that are more accurate and efficient after continuous learning and optimization.

[0154] This step collects user feedback information to generate a user feedback data set, analyzes user feedback with machine learning algorithms, identifies common problems and improvement directions, and dynamically adjusts the update strategy of the spectral database. Regularly update the spectral database, add new standard spectral data and sample information, and remove outdated and invalid data. Analyze a large amount of spectral data through data mining technology to discover potential sample characteristics and detection modes. Combined with the latest research results and industry standards, continuously optimize the spectral analysis algorithm to generate an optimized spectral database and detection mode.

[0155] Optionally, the customized detection result notification and personalized guidance suggestions described in step 104 are used to dynamically adjust the update strategy of the spectral database in combination with the machine learning algorithm, continuously optimize the spectral analysis algorithm, discover potential sample characteristics and detection modes through data mining technology, and generate an optimized spectral database and detection mode, including:

[0156] According to the customized test result notification and personalized guidance suggestions, user feedback information is collected to generate a user feedback data set;

[0157] Based on the user feedback data set, the user feedback information is analyzed in combination with a machine learning algorithm to identify common problems and improvement directions, dynamically adjust the update strategy of the spectral database to ensure the timeliness and accuracy of the database content, and generate an update strategy adjustment plan;

[0158] Using the update strategy adjustment scheme, regularly updating the spectral database, adding new standard spectral data and sample information, and removing obsolete and invalid data, to generate an optimized spectral database;

[0159] Based on the optimized spectral database, a large amount of spectral data is analyzed through data mining technology to discover potential sample characteristics and detection modes. Combined with the latest research results and industry standards, the spectral analysis algorithm is continuously optimized to generate optimized detection modes.

[0160] Optionally, the update strategy adjustment scheme is used to regularly update the spectral database, add new standard spectral data and sample information, and remove obsolete and invalid data to generate an optimized spectral database, including:

[0161] According to the update strategy adjustment plan, set the time interval and trigger conditions for regular updates and generate an update plan;

[0162] Based on the update plan, start an update process of the spectral database, collect the latest standard spectral data and sample information from multiple data sources, and generate a data set to be updated;

[0163] Using the data set to be updated, quality control is performed on the collected new data, and data cleaning and verification algorithms are used to remove noise and erroneous data to generate a high-quality new data set;

[0164] Based on the high-quality new data set, combined with the existing spectral database, data fusion technology is used to integrate the new data with the existing data, and the data validity evaluation algorithm is used to identify and remove outdated and invalid data to generate an optimized spectral database.

[0165] Optionally, according to the optimized spectral database, a large amount of spectral data is analyzed by data mining technology to discover potential sample characteristics and detection modes, and the spectral analysis algorithm is continuously optimized in combination with the latest research results and industry standards to generate an optimized detection mode, including:

[0166] Based on the optimized spectral database, a large amount of spectral data is deeply analyzed using data mining technology to identify potential characteristic patterns and abnormal phenomena in the sample and generate potential characteristic analysis results;

[0167] Based on the potential characteristic analysis results, combined with the latest research results and industry standards, the existing spectrum analysis algorithm is evaluated and improved, and new algorithm models and technical means are introduced to generate an improved spectrum analysis algorithm;

[0168] Using the improved spectral analysis algorithm, reanalyze the sample data in the optimized spectral database, verify the effect of the algorithm improvement, and generate an algorithm performance evaluation report;

[0169] According to the algorithm performance evaluation report, combined with user feedback and actual detection results, the spectral analysis algorithm is continuously optimized to form an optimized detection mode.

[0170] Assume that a remote spectrophotometric detection system based on a cloud platform has been running for a while and has collected a lot of user feedback information. The following is the specific implementation process:

[0171] First, the system automatically collects user feedback information, including evaluation of the accuracy of the test results, problems encountered during the test, and suggestions for improvement; for example, user A reported that the result of the glucose concentration in a certain test did not match the actual situation, and user B reported that the chart in the test report was not intuitive enough, etc.; these feedback information are organized into a user feedback dataset, which contains each user's feedback content, timestamp, and related metadata.

[0172] Secondly, based on the user feedback dataset, the user feedback information is analyzed in combination with a machine learning algorithm, common problems and improvement directions are identified, the update strategy of the spectral database is dynamically adjusted, the timeliness and accuracy of the database content are ensured, and an update strategy adjustment plan is generated: the user feedback dataset is analyzed using a machine learning algorithm (such as cluster analysis, text classification, etc.) to identify common problem types and improvement directions; for example, through cluster analysis, it is found that many user feedbacks focus on the accuracy of glucose concentration detection, while other users are concerned about the readability of the report; according to the analysis results, the update strategy of the spectral database is dynamically adjusted, such as giving priority to updating standard spectral data related to glucose detection, and improving the report generation template to improve readability; an update strategy adjustment plan is generated, including the update time interval, trigger conditions, and specific data sources.

[0173] Furthermore, the update strategy adjustment scheme is used to regularly update the spectral database, add new standard spectral data and sample information, and remove outdated and invalid data to generate an optimized spectral database: according to the update strategy adjustment scheme, the time interval for regular updates is set to once a month, and the trigger condition is that the update process is automatically triggered when a certain amount of new data accumulates; the update process of the spectral database is started, and the latest standard spectral data and sample information are collected from multiple data sources (such as laboratories, scientific research institutions, and public databases) to generate a data set to be updated; the quality of the collected new data is controlled, and data cleaning and verification algorithms are used to remove noise and erroneous data to generate a high-quality new data set; combined with the existing spectral database, the new data is integrated with the existing data using data fusion technology, and the data validity evaluation algorithm is used to identify and remove outdated and invalid data to generate an optimized spectral database.

[0174] Finally, according to the optimized spectral database, a large amount of spectral data is analyzed by data mining technology to discover potential sample characteristics and detection modes. In combination with the latest research results and industry standards, the spectral analysis algorithm is continuously optimized to generate an optimized detection mode: a large amount of spectral data in the optimized spectral database is deeply analyzed by data mining technology (such as association rule mining, cluster analysis, etc.) to identify potential characteristic patterns and abnormal phenomena in the samples; for example, through cluster analysis, it is found that the spectral signals at certain specific wavelengths are strongly correlated with the concentration of a certain component, or that some samples show abnormal spectral characteristics under specific conditions; based on these analysis results, combined with the latest research results and industry standards, the existing spectral analysis algorithm is evaluated and improved, new algorithm models and technical means are introduced, and an improved spectral analysis algorithm is generated; using the improved spectral analysis algorithm, the sample data in the optimized spectral database is re-analyzed to verify the effect of the algorithm improvement and generate an algorithm performance evaluation report; according to the algorithm performance evaluation report, combined with user feedback and actual detection effects, the spectral analysis algorithm is continuously optimized to form an optimized detection mode.

[0175] Through the above steps, the system can continuously learn and optimize, improve the accuracy and efficiency of detection, and better meet the needs of users.

[0176] This application considers that user feedback is a very important source of information in the maintenance and update process of the spectral database. By analyzing user feedback, common problems and improvement directions can be identified, so as to dynamically adjust the update strategy of the spectral database to ensure the timeliness and accuracy of the database content. Formula P k and U t Used to quantify this feedback information and generate update strategy adjustment plans.

[0177] Optionally, based on the user feedback data set, the user feedback information is analyzed in combination with a machine learning algorithm, common problems and improvement directions are identified, the update strategy of the spectral database is dynamically adjusted to ensure the timeliness and accuracy of the database content, and an update strategy adjustment plan is generated, including:

[0178] When calculating the recognition probability P of the kth common problem k Previously, the user feedback data was cleaned, sentiment analyzed, and keywords extracted to generate a relevance score r i and severity scores i , and count the frequency of problems f ik And record submission time t i ;

[0179]

[0180] Among them, P kis the recognition probability of the kth common problem; m is the total number of user feedback information; r i is the relevance score of the i-th user feedback information; f ik is the frequency of occurrence of the kth type of problem in the i-th user feedback information; λ is the weight factor of the logarithmic term, which is used to adjust the impact of the logarithmic term; s i is the severity score of the i-th user feedback information; η is the amplitude factor of the time impact item, which is used to adjust the impact of the time impact item; t i is the submission time of the i-th user feedback information, which is used to consider the time factor;

[0181] After calculating the recognition probability P of the kth common problem k Then, calculate the average P of all common problem recognition probabilities avg and standard deviation σ P , and combined with the time factor t to calculate the update strategy adjustment score U t ;

[0182] After calculating the characteristic spectrum signal intensity F i After that, the characteristic spectral signal is normalized and assigned a weight w i , calculate and generate the comprehensive evaluation score S j , reflecting the overall quality of the characteristic spectral signals of each component in the sample;

[0183]

[0184] Among them, U t is the update strategy adjustment score, which is used to decide whether the spectral database needs to be updated; t is the time since the last update; α is the amplitude factor of the Gaussian function, which is used to adjust the height of the Gaussian peak; P k is the recognition probability of the kth common problem; P avg is the average of all common problem recognition probabilities; σ P is the standard deviation of the probability of identifying common problems; β is the amplitude factor of the time decay term, which is used to adjust the influence of time; τ is the decay rate of the time decay term, which reflects the speed of the influence of time; γ is the coefficient of the weighted average term, which is used to adjust the influence of the weighted average term; n is the number of categories of common problems; w k is the weight of the kth type of common problem, reflecting the importance and urgency of the problem;

[0185] Calculate U t After that, set the threshold value, when U t When the threshold is exceeded, the update process of the spectral database is triggered. k and the weight of the common problem w k , identify common problem categories that need to be updated first and develop specific updates.

[0186] This formula is designed to dynamically adjust the update strategy of the spectral database through user feedback data to ensure the timeliness and accuracy of the database content. In this way, the system can automatically adjust the update strategy based on the user's actual feedback and time changes to ensure that the database content is always kept up to date and accurate, thereby improving user experience and data quality.

[0187] The following is a brief introduction to the design reasons of each sub-item of the formula:

[0188]

[0189] The reason for the design is to combine the relevance score r of user feedback i and the frequency f of a particular question k in these responses ik , to calculate the recognition probability of the kth common problem.

[0190] Logarithmic term Enhance the impact on the severity score, especially when the severity score is high, to increase the probability of recognition.

[0191] Time Influence Item Taking the time factor into consideration, the most recent feedback has a higher weight, ensuring the timeliness of the database content.

[0192] The following is a brief introduction to how to obtain the parameters of the formula:

[0193] Relevance score i Generate a relevance score for each user feedback through text cleaning, sentiment analysis, and keyword extraction; a severity score i Through sentiment analysis and keyword extraction, the severity score of each user feedback is generated; frequency statistics f ik By counting the frequency of each question in all user feedback; the time factor t i By recording the submission time of each user feedback; the weight factor λ of the logarithmic item is set to an initial value based on experience or literature, and adjusted through an optimization algorithm; the amplitude factor η of the time impact item is set to an initial value based on experience or literature, and adjusted through an optimization algorithm.

[0194] The following is a brief introduction to the design reasons of each sub-item of the formula:

[0195]

[0196] Gaussian function term Through Gaussian distribution, the deviation between the recognition probability and the average value is adjusted to ensure the rationality of the recognition probability.

[0197] Time decay term As time goes by, the need for updates is gradually increased to ensure that the database content is updated in a timely manner.

[0198] Weighted average Adjust the importance of different questions based on the weight of common questions to ensure that important questions are updated first.

[0199] The following is a brief introduction to how to obtain the parameters of the formula:

[0200] The amplitude factor α of the Gaussian function is set to an initial value based on experience or literature and adjusted through an optimization algorithm; the amplitude factor β of the time decay term is set to an initial value based on experience or literature and adjusted through an optimization algorithm; the decay rate τ of the time decay term is set to an initial value based on experience or literature and adjusted through an optimization algorithm; the coefficient γ of the weighted average term is set to an initial value based on experience or literature and adjusted through an optimization algorithm; the weight w of the common problem is k Set initial values ​​based on the importance and urgency of the issue, and adjust based on expert knowledge or historical data.

[0201] Assume that there is the following user feedback data:

[0202] The initial parameters are set as λ=0.5; η=0.3; α=1.0; β=0.2; τ=10; γ=0.5; w A =0.6; w B =0.7; w C =0.5;

[0203] Problem A: r1 = 0.8, r3 = 0.7; s1 = 0.7, s3 = 0.8; f1 A =1,f3 A =1

[0204] Problem B: r2 = 0.6, r5 = 0.9; s2 = 0.9, s5 = 0.7; f2 B =1,f5 B =1

[0205] Problem C: r4=0.5; s4=0.6; f 4C =1

[0206]

[0207] Assume that the time since the last update is t = 20 days.

[0208]

[0209] Identification probability P k: Question A and Question B have higher recognition probabilities (0.7076 and 0.7067 respectively), while Question C has lower recognition probability (0.2718). This indicates that Question A and Question B are more common questions in current user feedback. Update strategy to adjust score U t :U t =1.2448, assuming the threshold is 1.0, due to U t >1.0, triggering the update process of the spectral database. Priority update category: k and weight W k , question A and question B have higher weights and higher recognition probabilities, so the spectral data related to question A and question B should be updated first.

[0210] Through this embodiment, we can see how to generate an update strategy adjustment plan through user feedback data and formula calculation, and decide whether the spectral database needs to be updated and which problem categories should be updated first based on the results.

[0211] Figure 2 A schematic diagram of a remote spectrophotometric detection system based on a cloud platform is provided for the present application embodiment. Figure 2 As shown, the device comprises:

[0212] The receiving module 21 is used to synchronously receive spectral data from multiple remote spectrophotometers, and perform correction processing on the measurement differences between different instruments to obtain consistent sample spectral information;

[0213] The analysis module 22 is used to perform a detailed analysis of the concentration ratios of multiple components in the sample based on the consistent sample spectrum information using a high-order spectrum analysis algorithm combined with chemometrics technology, accelerate the calculation process through the distributed computing capability of the cloud platform, improve the detection accuracy and speed, and generate a detailed component analysis report;

[0214] The conversion module 23 is used to convert the key information in the component analysis report into an easy-to-understand text format according to the detailed component analysis report, provide personalized guidance and suggestions according to the specific needs of the user, and generate customized test result notifications and personalized guidance suggestions;

[0215] The adjustment module 24 is used to utilize the customized detection result notification and personalized guidance suggestions, combine with the machine learning algorithm to dynamically adjust the update strategy of the spectral database, continuously optimize the spectral analysis algorithm, discover potential sample characteristics and detection modes through data mining technology, and generate an optimized spectral database and detection mode.

[0216] Figure 2 The remote spectrophotometric detection system based on a cloud platform can perform Figure 1The implementation principle and technical effect of the remote spectrophotometric detection method based on a cloud platform described in the embodiment shown are not repeated here. The specific way in which each module and unit performs operations in the remote spectrophotometric detection system based on a cloud platform in the above embodiment has been described in detail in the embodiment of the method, and will not be elaborated here.

[0217] In one possible design, Figure 2 A remote spectrophotometric detection system based on a cloud platform of the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0218] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0219] The processing component 32 is used to: synchronously receive spectral data from multiple remote spectrophotometers, correct the measurement differences between different instruments, and obtain consistent sample spectral information; based on the consistent sample spectral information, use a high-order spectral analysis algorithm combined with chemometrics technology to perform a detailed analysis of the concentration ratios of multiple components in the sample, accelerate the calculation process through the distributed computing capabilities of the cloud platform, improve the detection accuracy and speed, and generate a detailed component analysis report; based on the detailed component analysis report, convert the key information in the component analysis report into an easy-to-understand text format, provide personalized guidance and suggestions based on the user's specific needs, and generate customized detection result notifications and personalized guidance suggestions; use the customized detection result notifications and personalized guidance suggestions, combined with machine learning algorithms to dynamically adjust the update strategy of the spectral database, continuously optimize the spectral analysis algorithm, discover potential sample characteristics and detection modes through data mining technology, and generate an optimized spectral database and detection mode.

[0220] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0221] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0222] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0223] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0224] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0225] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0226] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The illustrated embodiment is a remote spectrophotometric detection method based on a cloud platform.

[0227] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0228] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0229] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0230] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A remote spectrophotometric detection method based on a cloud platform, characterized in that: include: Synchronously receive spectral data from multiple remote spectrophotometers, correct the measurement differences between different instruments, and obtain consistent sample spectral information; Based on the consistent sample spectral information, a high-order spectral analysis algorithm combined with chemometrics technology is used to perform a detailed analysis of the concentration ratios of multiple components in the sample. The distributed computing capabilities of the cloud platform are used to accelerate the calculation process, improve the detection accuracy and speed, and generate a detailed component analysis report. According to the detailed component analysis report, key information in the component analysis report is converted into an easy-to-understand text format, personalized guidance and suggestions are provided according to the specific needs of the user, and customized test result notifications and personalized guidance suggestions are generated; Utilizing the customized test result notification and personalized guidance suggestions, combined with the machine learning algorithm, the update strategy of the spectral database is dynamically adjusted, the spectral analysis algorithm is continuously optimized, and potential sample characteristics and detection modes are discovered through data mining technology to generate an optimized spectral database and detection mode.

2. The method according to claim 1, characterized in that Based on the consistent sample spectral information, a high-order spectral analysis algorithm combined with chemometrics technology is used to perform a detailed analysis of the concentration ratios of multiple components in the sample. The distributed computing capabilities of the cloud platform are used to accelerate the calculation process, improve the detection accuracy and speed, and generate a detailed component analysis report, including: According to the consistent sample spectral information, a high-order spectral analysis algorithm is used to perform multi-dimensional analysis on the spectral data, extract characteristic spectral signals of various components in the sample, and generate a characteristic spectral signal set; Based on the characteristic spectral signal set, a multivariate regression model is established in combination with chemometrics technology, and the concentration ratios of multiple components in the sample are calculated by the model to generate preliminary component analysis results; Using the distributed computing capability of the cloud platform, the preliminary component analysis results are distributed to multiple computing nodes for parallel processing to generate optimized component analysis results; According to the optimized component analysis results, combined with the data visualization tool of the cloud platform, a detailed component analysis report is generated.

3. The method according to claim 2, characterized in that The method of performing multi-dimensional analysis on the spectral data according to the consistent sample spectral information by using a high-order spectral analysis algorithm, extracting characteristic spectral signals of various components in the sample, and generating a characteristic spectral signal set includes: In calculating the characteristic spectral signal intensity F i Previously, spectral data needed to be smoothed, baseline corrected, and normalized to ensure data accuracy and consistency; Among them, F i is the characteristic spectral signal intensity, corresponding to the characteristic spectral signal of the i-th component; I i is the original spectral intensity of the i-th component in the consistent sample spectral information; α is the amplitude factor of the Gaussian function; μ is the mean of the spectral intensity; σ is the standard deviation of the spectral intensity; β is the coefficient for adjusting the nonlinear term; γ is the nonlinear exponent; η is the amplitude factor of the sine function; λ is the wavelength of the sine function; θ is the amplitude factor of the exponential decay term; φ is the decay rate of the exponential decay term; After calculating the characteristic spectrum signal intensity F i After that, the characteristic spectral signal is normalized and assigned a weight w i , calculate and generate the comprehensive evaluation score S j , reflecting the overall quality of the characteristic spectral signals of each component in the sample; Among them, S j is a comprehensive evaluation score used to evaluate the overall quality of the characteristic spectral signal set of the jth sample; n is the number of components in the sample; w i is the weight of the i-th component, reflecting the importance of the component in the sample; F i is the characteristic spectrum signal intensity of the i-th component; max(F) is the maximum value of all characteristic spectrum signal intensities; δ is the power of the normalized characteristic spectrum signal intensity; F avg is the average value of the characteristic spectrum signal intensity; τ is the variance of the characteristic spectrum signal intensity; K is the coefficient of the logarithmic function; min(F) is the minimum value of all characteristic spectrum signal intensities; After calculating the comprehensive evaluation score S j After that, according to the comprehensive evaluation score S j The characteristic spectral signals are sorted and screened, cluster analysis is performed, characteristic spectral signal clusters are generated, and they are organized into structured data sets.

4. The method according to claim 2, characterized in that: The method of establishing a multivariate regression model based on the characteristic spectral signal set and combining chemometrics technology is used to calculate the concentration ratios of multiple components in the sample and generate preliminary component analysis results, including: According to the characteristic spectral signal set, a suitable chemometric method is selected to preprocess the characteristic spectral signal, remove noise and irrelevant variables, and obtain an optimized characteristic spectral signal; Based on the optimized characteristic spectral signal and in combination with known standard sample data, a multivariate regression model capable of establishing a mathematical relationship between the characteristic spectral signal and the concentration ratios of multiple components in the sample is established to generate a model training result; Using the model training results, the characteristic spectral signals of unknown samples are predicted, and the concentration ratios of multiple components in the samples are calculated through a multivariate regression model to generate prediction results; Based on the prediction results, combined with the confidence interval and error analysis of the model, preliminary component analysis results are generated.

5. The method according to claim 2, characterized in that: According to the optimized component analysis results, combined with the data visualization tool of the cloud platform, a detailed component analysis report is generated, including: According to the optimized component analysis results, the data processing function of the cloud platform is used to further count and summarize the analysis results, calculate the average concentration, standard deviation and confidence interval of each component, and generate statistical analysis data; Based on the statistical analysis data, combined with the data visualization tool of the cloud platform, an intuitive visualization chart is generated; Generate detailed report text using the visual chart in combination with detailed text description; Based on the detailed report text and the visual chart, combined with the user's historical test records and preference settings, the cloud platform's intelligent recommendation system provides the user with a customized report interpretation service and generates a detailed component analysis report.

6. The method according to claim 1, characterized in that The customized detection result notification and personalized guidance suggestions are used in combination with machine learning algorithms to dynamically adjust the update strategy of the spectral database, continuously optimize the spectral analysis algorithm, discover potential sample characteristics and detection modes through data mining technology, and generate optimized spectral databases and detection modes, including: According to the customized test result notification and personalized guidance suggestions, user feedback information is collected to generate a user feedback data set; Based on the user feedback data set, the user feedback information is analyzed in combination with a machine learning algorithm to identify common problems and improvement directions, dynamically adjust the update strategy of the spectral database to ensure the timeliness and accuracy of the database content, and generate an update strategy adjustment plan; Using the update strategy adjustment scheme, regularly updating the spectral database, adding new standard spectral data and sample information, and removing obsolete and invalid data, to generate an optimized spectral database; Based on the optimized spectral database, a large amount of spectral data is analyzed through data mining technology to discover potential sample characteristics and detection modes. Combined with the latest research results and industry standards, the spectral analysis algorithm is continuously optimized to generate optimized detection modes.

7. The method according to claim 6, characterized in that Based on the user feedback data set, the user feedback information is analyzed in combination with a machine learning algorithm, common problems and improvement directions are identified, the update strategy of the spectral database is dynamically adjusted to ensure the timeliness and accuracy of the database content, and an update strategy adjustment plan is generated, including: When calculating the recognition probability P of the kth common problem k Previously, the user feedback data was cleaned, sentiment analyzed, and keywords extracted to generate a relevance score r i and severity scores i , and count the frequency of problems f ik And record submission time t i ; Among them, P k is the recognition probability of the kth common problem; m is the total number of user feedback information; r i is the relevance score of the i-th user feedback information; f ik is the frequency of occurrence of the kth type of problem in the i-th user feedback information; λ is the weight factor of the logarithmic term, which is used to adjust the impact of the logarithmic term; s i is the severity score of the i-th user feedback information; η is the amplitude factor of the time impact item, which is used to adjust the impact of the time impact item; t i is the submission time of the i-th user feedback information, which is used to consider the time factor; After calculating the recognition probability P of the kth common problem k Then, calculate the average P of all common problem recognition probabilities avg and standard deviation σ P , and combined with the time factor t to calculate the update strategy adjustment score U t ; Among them, U t is the update strategy adjustment score, which is used to decide whether the spectral database needs to be updated; t is the time since the last update; α is the amplitude factor of the Gaussian function, which is used to adjust the height of the Gaussian peak; P k is the recognition probability of the kth common problem; P avg is the average of all common problem recognition probabilities; σ P is the standard deviation of the probability of identifying common problems; β is the amplitude factor of the time decay term, which is used to adjust the influence of time; τ is the decay rate of the time decay term, which reflects the speed of the influence of time; γ is the coefficient of the weighted average term, which is used to adjust the influence of the weighted average term; n is the number of categories of common problems; w k is the weight of the kth type of common problem, reflecting the importance and urgency of the problem; Calculate U t After that, set the threshold value, when U t When the threshold is exceeded, the update process of the spectral database is triggered. k and the weight of the common problem w k , identify common problem categories that need to be updated first and develop specific updates.

8. The method according to claim 6, characterized in that The update strategy adjustment scheme is used to regularly update the spectral database, add new standard spectral data and sample information, and remove obsolete and invalid data to generate an optimized spectral database, including: According to the update strategy adjustment plan, set the time interval and trigger conditions for regular updates and generate an update plan; Based on the update plan, start an update process of the spectral database, collect the latest standard spectral data and sample information from multiple data sources, and generate a data set to be updated; Using the data set to be updated, quality control is performed on the collected new data, and data cleaning and verification algorithms are used to remove noise and erroneous data to generate a high-quality new data set; Based on the high-quality new data set and combined with the existing spectral database, data fusion technology is used to integrate the new data with the existing data, and a data validity evaluation algorithm is used to identify and remove outdated and invalid data to generate an optimized spectral database.

9. The method according to claim 6, characterized in that According to the optimized spectral database, a large amount of spectral data is analyzed by data mining technology to discover potential sample characteristics and detection modes, and the spectral analysis algorithm is continuously optimized in combination with the latest research results and industry standards to generate optimized detection modes, including: Based on the optimized spectral database, a large amount of spectral data is deeply analyzed using data mining technology to identify potential characteristic patterns and abnormal phenomena in the sample and generate potential characteristic analysis results; Based on the potential characteristic analysis results, combined with the latest research results and industry standards, the existing spectrum analysis algorithm is evaluated and improved, and new algorithm models and technical means are introduced to generate an improved spectrum analysis algorithm; Using the improved spectral analysis algorithm, reanalyze the sample data in the optimized spectral database, verify the effect of the algorithm improvement, and generate an algorithm performance evaluation report; According to the algorithm performance evaluation report, combined with user feedback and actual detection results, the spectral analysis algorithm is continuously optimized to form an optimized detection mode.

10. A remote spectrophotometric detection system based on a cloud platform, characterized in that: include: A receiving module is used to synchronously receive spectral data from multiple remote spectrophotometers, correct the measurement differences between different instruments, and obtain consistent sample spectral information; An analysis module is used to perform a detailed analysis of the concentration ratios of multiple components in the sample based on the consistent sample spectral information using a high-order spectral analysis algorithm combined with chemometrics technology, accelerate the calculation process through the distributed computing capabilities of the cloud platform, improve the detection accuracy and speed, and generate a detailed component analysis report; A conversion module, for converting key information in the component analysis report into an easy-to-understand text format based on the detailed component analysis report, providing personalized guidance and suggestions according to the specific needs of the user, and generating customized test result notifications and personalized guidance suggestions; The adjustment module is used to utilize the customized detection result notification and personalized guidance suggestions, combine with the machine learning algorithm to dynamically adjust the update strategy of the spectral database, continuously optimize the spectral analysis algorithm, discover potential sample characteristics and detection modes through data mining technology, and generate an optimized spectral database and detection mode.

Images