Data processing method, device, equipment and storage medium of SaaS system

Abstract

The present application relates to the field of data processing technology, and discloses a data processing method, device, equipment and storage medium for a SaaS system. The method comprises: processing multiple data streams respectively through multiple data processing units in the SaaS system to obtain a first processing result for each data stream; performing parameter correction on the first processing results respectively to obtain multiple second processing results; performing data integration processing on the multiple second processing results to obtain target data integration information; performing data compression and view reconstruction on the target data integration information to obtain an initial data view model; performing feature analysis on the initial data view model to obtain a data view feature set; performing view optimization on the data view feature set and the initial data view model to obtain the target data view model of the SaaS system. The present application improves the accuracy of data processing in the SaaS system.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a data processing method, device, equipment and storage medium for a SaaS system. Background Art

[0002] As an important part of cloud computing, SaaS system provides enterprises with flexible and scalable software solutions. This model allows users to access applications through the Internet without the need to maintain and update software locally, greatly reducing IT expenditures and operating costs. However, with the dramatic increase in data volume and the diversification of business needs, existing SaaS systems face the dual challenges of data processing efficiency and processing quality. Data is constantly pouring in from multiple sources, and how to quickly and accurately process this data has become the key to improving service quality and user experience.

[0003] In addition, data integration processing is particularly critical in SaaS systems. Data from different data sources and formats need to be integrated to provide a unified view and in-depth business insights. Existing methods often lack effective mechanisms to process and integrate these heterogeneous data, resulting in serious data silos and affecting the real-time and accuracy of data. Therefore, how to optimize the data integration process and improve data utilization and processing speed is one of the problems that the current SaaS system urgently needs to solve. Furthermore, with the development of big data technology, it has become particularly important to effectively compress and extract features from data. Although data compression can reduce storage space and improve processing efficiency, inappropriate compression may lead to information loss and affect the quality and accuracy of data analysis. Summary of the invention

[0004] The present application provides a data processing method, apparatus, device and storage medium for a SaaS system, which are used to improve the accuracy of data processing of the SaaS system.

[0005] In a first aspect, the present application provides a data processing method for a SaaS system, the data processing method for the SaaS system comprising:

[0006] Processing the multiple data streams respectively by the multiple data processing units in the SaaS system to obtain a first processing result of each data stream;

[0007] Performing parameter correction on the first processing results respectively to obtain a plurality of second processing results;

[0008] Performing data integration processing on the plurality of second processing results to obtain target data integration information;

[0009] Performing data compression and view reconstruction on the target data integration information to obtain an initial data view model;

[0010] Performing feature analysis on the initial data view model to obtain a data view feature set;

[0011] View optimization is performed on the data view feature set and the initial data view model to obtain a target data view model of the SaaS system.

[0012] In a second aspect, the present application provides a data processing device for a SaaS system, the data processing device for the SaaS system comprising:

[0013] A processing module, used to process the multiple data streams respectively through the multiple data processing units in the SaaS system to obtain a first processing result of each data stream;

[0014] A correction module, used for respectively performing parameter correction on the first processing results to obtain a plurality of second processing results;

[0015] An integration module, used for performing data integration processing on the plurality of second processing results to obtain target data integration information;

[0016] A reconstruction module, used for performing data compression and view reconstruction on the target data integration information to obtain an initial data view model;

[0017] An analysis module, used for performing feature analysis on the initial data view model to obtain a data view feature set;

[0018] The optimization module is used to perform view optimization on the data view feature set and the initial data view model to obtain a target data view model of the SaaS system.

[0019] In a third aspect, the present application provides a data processing device for a SaaS system, comprising: a memory and at least one processor, wherein instructions are stored in the memory; the at least one processor calls the instructions in the memory so that the data processing device of the SaaS system executes the above-mentioned data processing method of the SaaS system.

[0020] A fourth aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the data processing method of the SaaS system described above.

[0021] In the technical solution provided by the present application, by segmenting and parameter correction of the processing units of multiple data streams, the accuracy and efficiency of data processing can be significantly improved. Each data processing unit corresponds to a specific data task, making the processing process more specialized and refined, thereby reducing errors and improving processing speed. This segmentation and correction method also allows the system to dynamically adjust the parameters of each processing unit to meet the needs of different types of data, which is particularly important when processing large-scale heterogeneous data. By integrating and optimizing the processed data, a unified data view can be generated, greatly enhancing the visualization and availability of the data. This integrated view not only provides a clearer and more comprehensive data display for the end user, but also provides a solid foundation for further data analysis and decision support. The optimized data integration results ensure the integrity and consistency of information and reduce data redundancy and inconsistency. By applying data compression and view reconstruction techniques, the cost of data storage is effectively reduced and the speed of data processing is improved. Data compression reduces the required storage space, while feature extraction and dimensionality reduction techniques ensure the minimum loss of data information during the compression process. In addition, feature analysis and view optimization further enhance the predictive power and analytical depth of the data model, providing the SaaS system with a more accurate and insightful data view, supporting more efficient business decisions, and thus improving the accuracy of data processing in the SaaS system. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0023] Figure 1 This is a schematic diagram of an embodiment of a data processing method of a SaaS system in an embodiment of the present application;

[0024] Figure 2 Schematic diagram of an embodiment of a data processing device of a SaaS system in an embodiment of the present application. DETAILED DESCRIPTION

[0025] The embodiments of the present application provide a data processing method, apparatus, device and storage medium for a SaaS system. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0026] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 , an embodiment of the data processing method of the SaaS system in the embodiment of the present application includes:

[0027] Step S101: Process multiple data streams respectively through multiple data processing units in the SaaS system to obtain a first processing result of each data stream;

[0028] It is understandable that the execution subject of the present application may be a data processing device of the SaaS system, or a terminal or a server, which is not limited here. The present application embodiment is described by taking the server as the execution subject as an example.

[0029] Specifically, a data processing architecture that can support parallel processing is constructed. This architecture usually relies on a distributed computing environment, in which each data processing unit can be an independent virtual machine or a containerized service, and these units are distributed on the cloud infrastructure to improve processing efficiency and system scalability. For each data stream, a corresponding processing unit is assigned according to the source and nature of the data. For example, real-time data streams may require faster processing speeds and real-time analysis capabilities, while historical data streams may focus more on deep data mining and long-term trend analysis. After receiving the data stream, each processing unit performs data preprocessing, including data cleaning, denoising, format standardization and other steps to ensure that the data quality meets the requirements of subsequent processing. According to the specific application scenario, appropriate data processing algorithms are selected, such as data classification, cluster analysis, time series analysis, etc. These algorithms can specifically solve the analysis requirements of different types of data streams. Each data processing unit outputs a processing result, and the first processing result usually includes a preliminary analysis result of the data stream, such as a statistical summary, key indicator extraction, etc.

[0030] Step S102, performing parameter correction on the first processing results respectively to obtain a plurality of second processing results;

[0031] Specifically, it is defined that each data processing unit corresponds to a data processing channel, and specific parameters of each channel are obtained, which may include data transmission rate, processing delay, error rate, etc. The channel parameters are segmented, and each parameter is decomposed into several continuous or non-continuous segments, each segment representing the performance of the parameter under different operating conditions or time windows. Segmentation can help understand the changing rules of parameters under different conditions. A linear regression model is constructed for each data processing channel based on multiple parameter segmented data, and the performance of each data channel in future operations is predicted based on the historical data of the parameters through statistical learning methods. Linear regression is a commonly used prediction technology, which is suitable for processing situations where there is a linear relationship between parameters and performance. The performance of multiple data processing channels is predicted based on the linear regression model to obtain performance prediction data for each channel. The actual performance data of each channel in actual operation is obtained, which is usually obtained through real-time monitoring and historical performance records. The performance deviation of each channel is calculated by comparing the predicted data with the actual data. The first processing result of each data stream is corrected accordingly based on the deviation data. The specific performance deviation of each data processing channel is considered to ensure that the accuracy and reliability of the second processing result are significantly improved. Through a comprehensive parameter correction method, the SaaS system can optimize its data processing flow, improve data processing quality and efficiency, and ultimately obtain multiple second processing results, thereby enhancing the data processing capabilities and business performance of the entire system.

[0032] Step S103, performing data integration processing on the plurality of second processing results to obtain target data integration information;

[0033] Specifically, multiple second processing results are aligned and performance balanced to ensure that the results of different data sources can be compared and integrated under a unified standard. Data alignment mainly solves the consistency problem of data format and timestamp, ensuring that all data are processed and analyzed under the same conditions. Performance balancing adjusts the processing rate and quality of each data stream through an algorithm to avoid the data of high-performance channels overwhelmingly affecting the overall result, ensuring the fairness and balance of data processing. Through data fusion technology, multiple standard processing results are integrated according to the data stream sequence of each data processing channel. For example, time series analysis methods or event-driven data merging strategies are used to ensure the logical continuity and integrity of data. Integration relies on the timestamp of the data and considers the importance and priority of the data to generate initial data integration information. By evaluating the contribution and performance of each channel, performance weights are assigned to each data processing channel. The basis for weight allocation may include factors such as data accuracy, processing speed, and update frequency. Weighted analysis is performed on the initial data integration information according to the channel weight value. The influence of each channel in data integration is adjusted according to its weight, maximizing the advantages of each channel while suppressing the negative impact of those low-performance or unreliable data channels, ensuring that the accuracy and reliability of the target data integration information are significantly improved, and obtaining the target data integration information.

[0034] Step S104: compress the target data integration information and reconstruct the view to obtain an initial data view model;

[0035] Specifically, the data average of each data processing channel in the target data integration information is calculated respectively to evaluate the relative position and importance of each data channel in the overall data set. Through the average value, the data is normalized, and the data scale is adjusted to make its mean zero and variance one, eliminating the influence of different data magnitudes and units. The covariance matrix of the normalized data view is calculated, and the covariance matrix can reveal the correlation between different data dimensions. By performing eigenvalue decomposition on the covariance matrix, the eigenvalues ​​and corresponding eigenvectors are obtained. These eigenvectors represent the main change direction of the data, and the size of the eigenvalue indicates the degree of data variation in each main change direction. According to the size of the eigenvalue, the eigenvectors corresponding to the largest eigenvalues ​​are selected as the principal component features. These principal component features reflect the most important structural information in the data and are the key to dimensionality reduction mapping and data compression. By applying principal component analysis, the dimension of the data is reduced while retaining the most important information in the data, and the target compressed data is obtained. Dimensionality reduction can reduce the burden of data processing and storage, and improve the efficiency of data processing and the clarity of visualization. The compressed data is reconstructed and data features and model features are extracted from the target compressed data, including statistical characteristics, pattern recognition and potential trends of the data. Based on these features and model features, an initial data view model is constructed, which has low complexity and can provide intuitive understanding and analysis of the data.

[0036] Extract data response features from the target data integration information, identify and analyze key behaviors and response patterns in the data, for example, identify outliers, trend lines and periodic changes in the data through statistical analysis and machine learning techniques. At the same time, extract model space distribution features from the target data integration information. Analyze the distribution characteristics and structural configuration of the data, for example, determine the cluster center and distribution boundary of the data points through cluster analysis and density estimation. These model features are key factors in optimizing the data model and improving the data explanatory power. Reconstruct the compressed target data to create an initial data model. The data model converts the compressed data points into a new data representation through specific mathematical and statistical methods, such as principal component regression or multidimensional scaling analysis, in which each data point integrates data information from different channels. Optimize the initial data model using the extracted data features and model features. The optimization process includes adjusting model parameters, improving data fit and enhancing the model's predictive ability to ensure that the model can effectively reflect the core characteristics and key information of the data. Through optimization, the initial data model is transformed into an initial data view model.

[0037] Step S105: Perform feature analysis on the initial data view model to obtain a data view feature set;

[0038] Specifically, high-dimensional data feature extraction is performed on the initial data view model, and features that can represent the core attributes and structure of the model are extracted from the initial data view model. Data analysis techniques, such as multidimensional scaling analysis or high-dimensional statistical modeling, are used to identify the most influential features from a large number of data variables and capture those data points that show significant changes or trends in multidimensional space. Data relationship feature extraction is performed on the initial data view model to obtain the intrinsic connections and interdependencies between data points. Model data relationship features are obtained by analyzing how data points interact with each other and how they influence each other in the data model. These relationship features may include distance metrics between data points, association rules, or node connectivity in a graph structure. A data view feature set of the initial data view model is created based on the model data features and the model data relationship features. The feature set reflects the multidimensional characteristics of the data and includes the relationships between data elements.

[0039] Step S106: Optimize the data view feature set and the initial data view model to obtain a target data view model of the SaaS system.

[0040] Specifically, the data view feature set and the initial data view model are parsed to extract and identify specific data feature parsing data and data model parsing data. The data feature parsing data mainly includes key features identified from the data view feature set, such as major data trends, patterns, anomalies, and potential data relationships. The data model parsing data includes information on the structure and functional attributes of the initial data view model, such as the dimensions of the data model, the correlation between data, and its adaptability to specific query and analysis tasks. The parsed data features and model data are subjected to view mapping and view enhancement processing. View mapping is a process of converting abstract data features into visual representations, mapping different data view features to one or more visual elements, such as color, shape, or size, so that users can intuitively understand the meaning and relationship of the data. View enhancement processing can improve the representation of visual elements, such as by adjusting the layout, optimizing the interactive design, or applying advanced graphics rendering technology to enhance the expressiveness of the view and user experience. Through effective view mapping and enhancement of data feature parsing data and data model parsing data, the SaaS system can generate an optimized target data view model.

[0041] In the embodiment of the present application, by segmenting and parameter correction of the processing units of multiple data streams, the accuracy and efficiency of data processing can be significantly improved. Each data processing unit corresponds to a specific data task, making the processing process more specialized and refined, thereby reducing errors and improving processing speed. This segmentation and correction method also allows the system to dynamically adjust the parameters of each processing unit to meet the needs of different types of data, which is particularly important when processing large-scale heterogeneous data. By integrating and optimizing the processed data, a unified data view can be generated, which greatly enhances the visualization and availability of the data. This integrated view not only provides a clearer and more comprehensive data display for the end user, but also provides a solid foundation for further data analysis and decision support. The optimized data integration results ensure the integrity and consistency of information and reduce the problems of data redundancy and inconsistency. By applying the technology of data compression and view reconstruction, the data storage cost is effectively reduced and the speed of data processing is improved. Data compression reduces the required storage space, while feature extraction and dimensionality reduction techniques ensure the minimum loss of data information during the compression process. In addition, feature analysis and view optimization further enhance the predictive power and analytical depth of the data model, providing the SaaS system with a more accurate and insightful data view, supporting more efficient business decisions, and thus improving the accuracy of data processing in the SaaS system.

[0042] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0043] (1) Define that each data processing unit corresponds to a data processing channel, and obtain the channel parameters of each data processing channel;

[0044] (2) Segment processing is performed on the channel parameters to obtain multiple parameter segments corresponding to each channel parameter;

[0045] (3) Constructing a linear regression model for each data processing channel according to multiple parameter segments;

[0046] (4) Predicting the performance of multiple data processing channels based on the linear regression model to obtain channel performance prediction data for each data processing channel;

[0047] (5) obtaining the actual channel performance data of the first processing result, and performing deviation calculation on the channel performance prediction data and the actual channel performance data to obtain a plurality of deviation data;

[0048] (6) Parameter correction is performed on the first processing result of each data stream according to the plurality of deviation data to obtain a plurality of second processing results.

[0049] Specifically, define each data processing unit to correspond to a data processing channel, and obtain the channel parameters of each data processing channel, which may include key performance indicators such as data throughput, processing time, and error rate. Perform segmented processing on the channel parameters, and decompose the continuous parameter values ​​into multiple interval segments. For example, assume that a video SaaS platform has multiple data processing channels, and each channel is responsible for rendering and processing videos with different resolutions. Channel parameters such as processing time are segmented according to the resolution, duration, and frame rate of the video. For example, the processing time is segmented according to the video length of 0-10 minutes, 10-30 minutes, and more than 30 minutes, and each segment is subdivided according to the video resolution and frame rate. A linear regression model is constructed for the data in each segment, and the model predicts the performance of each data processing channel in future operations based on historical data. Linear regression is a prediction technique that assumes that there is a linear relationship between variables and is suitable for predicting the relationship between a response variable (such as processing time) and one or more explanatory variables (such as video duration and resolution). For example, the model of each channel will be able to predict the processing time based on the input video characteristics (duration, resolution, frame rate). Use a linear regression model to predict the performance of each data processing channel and generate performance prediction data for each channel. Obtain the actual channel performance data of the channel in actual operation, which usually comes from the system's monitoring tools and log records. After the performance prediction, compare the predicted data with the actual operation data to calculate the performance prediction deviation. Perform parameter correction on the first processing result of each data stream based on the deviation data. Consider the specific performance deviation of each data processing channel to ensure that the accuracy and reliability of the second processing result are improved. For example, if it is found that the processing time is generally less than the prediction, it may be necessary to adjust the parameters of the regression model or re-analyze the processing strategy to ensure the processing quality and efficiency of each video.

[0050] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0051] (1) performing data alignment and performance balancing on the multiple second processing results to obtain multiple standard processing results;

[0052] (2) Integrate multiple standard processing results according to the data flow sequence of multiple data processing channels to generate initial data integration information;

[0053] (3) Allocate performance weights to multiple data processing channels respectively to obtain a channel weight value for each data processing channel;

[0054] (4) Perform weighted analysis on the initial data integration information according to the channel weight value to obtain the target data integration information.

[0055] Specifically, multiple second processing results from different data processing channels are data aligned and performance balanced. Data alignment ensures that all data streams are consistent in time or space, and usually involves adjusting timestamps, synchronizing data records, or standardizing data formats. Performance balancing involves adjusting resource allocation during data processing to ensure that no single data source or channel affects the quality of the overall result due to overload or insufficient performance. For example, if a data processing channel slows down due to processing high-resolution video, it may be necessary to increase computing resources or optimize algorithms to reduce processing delays. The aligned and performance-balanced standard processing results are integrated according to the data stream sequence of each data processing channel. During the integration process, the data should be arranged in a chronological or logical order, and the content and importance of the data should be considered. Performance weights are assigned to each data processing channel based on past performance records, data accuracy, and channel reliability. For example, if a channel frequently provides high-accuracy data, it may be given a higher weight. The weight reflects the importance and influence of each channel in the overall data integration process. The initial data integration information is weighted and analyzed according to the channel weight value. The data contribution of each channel is weighted using the weight value to ensure that the data analysis results can reflect the most reliable and influential information.

[0056] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0057] (1) Calculate the data average of each data processing channel in the target data integration information respectively;

[0058] (2) Normalizing the target data integration information according to the data average value to obtain a normalized data view;

[0059] (3) Calculate the covariance matrix corresponding to the normalized data view and perform eigenvalue decomposition on the covariance matrix to obtain the eigenvalues ​​and corresponding eigenvectors;

[0060] (4) According to the size of the eigenvalue, the principal component features are selected, and the target data integration information is mapped and compressed according to the principal component features to obtain the target compressed data;

[0061] (5) Extract data features and model features from the target data integration information respectively to obtain data features and model features, and reconstruct the view of the target compressed data to obtain an initial data view model.

[0062] Specifically, the data average of each data processing channel in the target data integration information is calculated respectively. The average value helps to identify the average performance or state of each channel within a certain time range. The target data integration information is normalized according to the data average value to obtain a normalized data view. By adjusting the range and scale of the data, data of different scales and dimensions can be compared and analyzed under the same standard. The covariance matrix corresponding to the normalized data view is calculated. The covariance matrix provides a measure of the change relationship between different variables, which can help identify which variables have a strong mutual influence. Eigenvalue decomposition is performed to extract eigenvalues ​​and corresponding eigenvectors from the covariance matrix. The eigenvalue reflects the variance size in each direction of the data, while the eigenvector represents these directions. By analyzing the size of the eigenvalue, it is determined which directions have the largest data variation, and these directions are usually regarded as the main components in the data. According to the main component characteristics, the target data integration information is subjected to dimensionality reduction mapping and data compression. Dimensionality reduction mapping converts multidimensional data into a lower dimensional space, reduces the complexity of the data set, and retains the most important information as much as possible. For example, when processing large data sets containing hundreds of variables, dimensionality reduction can significantly improve the efficiency of data processing and analysis. The reduced-dimensional data can be compressed to reduce the storage space required, while facilitating fast transmission and processing. The reduced-dimensional and compressed data can be reconstructed. Reconstructing the view is to create a data representation that is easier to understand and analyze, and to express the compressed data through various data visualization techniques. At the same time, data features and model features are extracted, including statistical properties, key indicators, and patterns of the data, which help to deeply understand the essential characteristics and trends of the data.

[0063] In a specific embodiment, the execution step extracts data features and model features from the target data integration information to obtain data features and model features, and reconstructs the view of the target compressed data to obtain the initial data view model, which may specifically include the following steps:

[0064] (1) Extracting data response features from the target data integration information to obtain data features, and extracting model space distribution features from the target data integration information to obtain model features;

[0065] (2) reconstructing the target compressed data to obtain an initial data model, wherein each data point in the initial data model represents data that combines data information from different channels;

[0066] (3) Based on the data characteristics and model characteristics, the initial data model is optimized to obtain the initial data view model.

[0067] Specifically, data response features are extracted from the target data integration information. Data response features reflect how data responds to changes in certain variables or environments. Model space distribution features are extracted from the target data integration information to analyze the distribution and structure of data in multidimensional space. These features help reveal hidden patterns and associations in the data, such as clustering tendencies, the presence of outliers, and correlations between variables. Data reconstruction is performed on the target compressed data to form an initial data model. In the data compression process, the dimensions of the data are reduced to remove redundancy and retain the most important information through techniques such as principal component analysis or autoencoders. The compressed data is reconstructed into a new data model, in which each data point not only reflects the information of a single data channel, but also combines comprehensive information from multiple different data channels. Based on the data features and model features, the initial data model is optimized. Model optimization may include algorithm adjustment, parameter fine-tuning, and retraining. In the optimization process, the data model is adjusted through repeated iterations until the expected performance standards are achieved. Finally, the initial data view model is obtained.

[0068] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0069] (1) Extract high-dimensional data features from the initial data view model to obtain model data features;

[0070] (2) Extracting data relationship features from the initial data view model to obtain model data relationship features;

[0071] (3) Create a data view feature set of the initial data view model based on the model data features and the model data relationship features.

[0072] Specifically, the initial data view model is subjected to high-dimensional data feature extraction, key information and performance indicators are identified from complex data sets, and various algorithms and techniques are applied to identify the main patterns and trends in the data to obtain model data features. When extracting high-dimensional data features, statistical analysis methods such as principal component analysis (PCA) or factor analysis are usually used to reduce the complexity of the data while retaining the most important features for explaining data variation. The initial data view model is subjected to data relationship feature extraction to determine the interdependence and connection between different variables in the data. By constructing and analyzing correlation matrices, covariance matrices or more advanced graph theory techniques, the internal structure and dynamics of the data are revealed, the hidden connections and patterns in the data are obtained, and the interactions between variables and their impact on the overall data behavior are understood. A data view feature set of the initial data view model is created based on the model data features and the model data relationship features. This feature set integrates the complex relationships between single feature points extracted from the data and data points, providing a multi-dimensional view for further data analysis and model optimization. Through the comprehensive feature extraction and analysis process, the system can better design and optimize data-driven models and improve the prediction accuracy and explanatory power of the model. For example, when optimizing algorithm parameters, the information extracted from data relationship features is used to adjust the model to better reflect these relationships, thereby achieving better performance in various application scenarios.

[0073] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0074] (1) Parse the data view feature set and the initial data view model to obtain data feature parsing data and data model parsing data;

[0075] (2) Perform view mapping and view enhancement on the data feature analysis data and the data model analysis data to obtain the target data view model of the SaaS system.

[0076] Specifically, the data view feature set and the initial data view model are parsed to analyze the high-dimensional data features and data relationship features that have been identified and extracted, and how they are represented in the initial data view model. This includes evaluating the statistical properties of each feature, the relationship between each other, and their impact on the entire data set. Convert complex data and relationships into a more understandable format to obtain data feature parsing data and data model parsing data. Data feature parsing data may include the distribution, central tendency (such as mean, median) and degree of dispersion (such as variance, standard deviation) of each feature, while data model parsing data focuses on the structure of the model, such as network relationships, hierarchical or clustered distributions between data. Perform view mapping and view enhancement on data feature parsing data and data model parsing data to convert the analysis results into intuitive and easy-to-understand visual representations. View mapping involves selecting appropriate visualization tools and techniques to display data. This may include using charts, graphs, heat maps, etc. to display data distribution and trends, or using network diagrams and tree diagrams to reveal structural relationships in data models. View enhancement processing improves the clarity and interactivity of these visual representations, ensuring that users can not only see the static view of the data, but also explore the detailed level of the data through interactive operations (such as clicking, dragging, zooming, etc.). Ultimately, the target data view model of the SaaS system is obtained.

[0077] The above describes the data processing method of the SaaS system in the embodiment of the present application. The following describes the data processing device of the SaaS system in the embodiment of the present application. Figure 2 In one embodiment of the present application, a data processing device of the SaaS system includes:

[0078] The processing module 201 is used to process multiple data streams respectively through multiple data processing units in the SaaS system to obtain a first processing result of each data stream;

[0079] A correction module 202, used to perform parameter correction on the first processing results respectively to obtain a plurality of second processing results;

[0080] An integration module 203 is used to perform data integration processing on the plurality of second processing results to obtain target data integration information;

[0081] The reconstruction module 204 is used to perform data compression and view reconstruction on the target data integration information to obtain an initial data view model;

[0082] An analysis module 205 is used to perform feature analysis on the initial data view model to obtain a data view feature set;

[0083] The optimization module 206 is used to perform view optimization on the data view feature set and the initial data view model to obtain a target data view model of the SaaS system.

[0084] Through the synergy of the above components, the accuracy and efficiency of data processing can be significantly improved by segmenting and parameter correction of the processing units of multiple data streams. Each data processing unit corresponds to a specific data task, making the processing process more specialized and refined, thereby reducing errors and improving processing speed. This segmentation and correction method also allows the system to dynamically adjust the parameters of each processing unit to meet the needs of different types of data, which is particularly important when processing large-scale heterogeneous data. By integrating and optimizing the processed data, a unified data view can be generated, which greatly enhances the visualization and usability of the data. This integrated view not only provides a clearer and more comprehensive data display for the end user, but also provides a solid foundation for further data analysis and decision support. The optimized data integration results ensure the integrity and consistency of information and reduce the problems of data redundancy and inconsistency. By applying data compression and view reconstruction techniques, the cost of data storage is effectively reduced and the speed of data processing is improved. Data compression reduces the required storage space, while feature extraction and dimensionality reduction techniques ensure minimal loss of data information during the compression process. In addition, feature analysis and view optimization further enhance the predictive power and analytical depth of the data model, providing the SaaS system with a more accurate and insightful data view, supporting more efficient business decisions, and thus improving the accuracy of data processing in the SaaS system.

[0085] The present application also provides a data processing device for a SaaS system, the data processing device for the SaaS system comprising a memory and a processor, the memory storing computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor executes the steps of the data processing method for the SaaS system in the above-mentioned embodiments.

[0086] The present application also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the data processing method of the SaaS system.

[0087] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0088] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0089] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, 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 data processing method for a SaaS system, characterized in that: The data processing method of the SaaS system includes: Processing the multiple data streams respectively by the multiple data processing units in the SaaS system to obtain a first processing result of each data stream; Performing parameter correction on the first processing results respectively to obtain a plurality of second processing results; Performing data integration processing on the multiple second processing results to obtain target data integration information; specifically including: performing data alignment and performance balancing on the multiple second processing results to obtain multiple standard processing results; integrating the multiple standard processing results according to the data flow sequence of multiple data processing channels to generate initial data integration information; performing performance weight allocation on the multiple data processing channels respectively to obtain a channel weight value of each data processing channel; performing weighted analysis on the initial data integration information according to the channel weight value to obtain target data integration information; The target data integration information is compressed and reconstructed to obtain an initial data view model; specifically comprising: calculating the data average value of each data processing channel in the target data integration information respectively; normalizing the target data integration information according to the data average value to obtain a normalized data view; calculating the covariance matrix corresponding to the normalized data view, and performing eigenvalue decomposition on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors; selecting principal component features according to the size of the eigenvalues, and performing dimensionality reduction mapping and data compression on the target data integration information according to the principal component features to obtain target compressed data; extracting data response features from the target data integration information to obtain data features, and extracting model space distribution features from the target data integration information to obtain model features; reconstructing the target compressed data to obtain an initial data model, wherein each data point in the initial data model represents data combining data information of different channels; optimizing the initial data model according to the data features and the model features to obtain an initial data view model; Performing feature analysis on the initial data view model to obtain a data view feature set; Performing view optimization on the data view feature set and the initial data view model to obtain a target data view model of the SaaS system; specifically comprising: parsing the data view feature set and the initial data view model to obtain data feature parsing data and data model parsing data; performing view mapping and view enhancement processing on the data feature parsing data and the data model parsing data to obtain a target data view model of the SaaS system.

2. The data processing method of the SaaS system according to claim 1, characterized in that: The performing parameter correction on the first processing results respectively to obtain a plurality of second processing results comprises: Define that each data processing unit corresponds to a data processing channel, and obtain channel parameters of each data processing channel; Segment processing is performed on the channel parameters to obtain a plurality of parameter segments corresponding to each channel parameter; Constructing a linear regression model for each data processing channel according to the multiple parameter segments; Performing performance prediction on the multiple data processing channels according to the linear regression model to obtain channel performance prediction data of each data processing channel; Acquire the channel performance real data of the first processing result, and perform deviation calculation on the channel performance prediction data and the channel performance real data to obtain a plurality of deviation data; Parameter correction is performed on the first processing result of each data stream according to the multiple deviation data to obtain multiple second processing results.

3. The data processing method of the SaaS system according to claim 1, characterized in that: The performing feature analysis on the initial data view model to obtain a data view feature set includes: Extracting high-dimensional data features from the initial data view model to obtain model data features; Extracting data relationship features from the initial data view model to obtain model data relationship features; A data view feature set of the initial data view model is created according to the model data features and the model data relationship features.

4. A data processing device for a SaaS system, characterized in that: The data processing device of the SaaS system includes: A processing module, used to process the multiple data streams respectively through the multiple data processing units in the SaaS system to obtain a first processing result of each data stream; A correction module, used for respectively performing parameter correction on the first processing results to obtain a plurality of second processing results; The integration module is used to perform data integration processing on the multiple second processing results to obtain target data integration information; specifically comprising: performing data alignment and performance balancing on the multiple second processing results to obtain multiple standard processing results; integrating the multiple standard processing results according to the data flow sequence of multiple data processing channels to generate initial data integration information; performing performance weight allocation on the multiple data processing channels respectively to obtain a channel weight value of each data processing channel; performing weighted analysis on the initial data integration information according to the channel weight value to obtain target data integration information; A reconstruction module is used to perform data compression and view reconstruction on the target data integration information to obtain an initial data view model; specifically comprising: respectively calculating the data average value of each data processing channel in the target data integration information; normalizing the target data integration information according to the data average value to obtain a normalized data view; calculating the covariance matrix corresponding to the normalized data view, and performing eigenvalue decomposition on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors; selecting principal component features according to the size of the eigenvalues, and performing dimensionality reduction mapping and data compression on the target data integration information according to the principal component features to obtain target compressed data; extracting data response features from the target data integration information to obtain data features, and extracting model space distribution features from the target data integration information to obtain model features; reconstructing the target compressed data to obtain an initial data model, wherein each data point in the initial data model represents data combining data information of different channels; optimizing the initial data model according to the data features and the model features to obtain an initial data view model; An analysis module, used for performing feature analysis on the initial data view model to obtain a data view feature set; The optimization module is used to perform view optimization on the data view feature set and the initial data view model to obtain the target data view model of the SaaS system; specifically comprising: parsing the data view feature set and the initial data view model to obtain data feature parsing data and data model parsing data; performing view mapping and view enhancement processing on the data feature parsing data and the data model parsing data to obtain the target data view model of the SaaS system.

5. A data processing device for a SaaS system, characterized in that: The data processing device of the SaaS system includes: a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instruction in the memory so that the data processing device of the SaaS system executes the data processing method of the SaaS system according to any one of claims 1 to 3.

6. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the data processing method of the SaaS system according to any one of claims 1 to 3 is implemented.

Images