Cross-source data collaborative analysis method and system based on AI model

Abstract

The present invention relates to the field of electronic digital data processing technology, specifically to a cross-source data collaborative analysis method and system based on an AI model, the method comprising: regularly conducting a value assessment on each sales channel, screening out platforms with data collection value, marking them as each newly added platform, and jointly marking them with the original data collection platform as each data collection platform; developing a real-time monitoring algorithm for data changes, analyzing the changing trend of smart home product retail data of each data collection platform, automatically adjusting the data collection frequency for data collection, and obtaining multi-source data; using a distributed network architecture for data transmission, real-time monitoring of data transmission status information, determining whether data transmission is abnormal, and providing early warning feedback. The present invention optimizes data sources, collection, and transmission by providing a cross-source data collaborative analysis method and system based on an AI model, and provides more comprehensive and accurate information for the AI ​​model to realize cross-source data collaborative analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic digital data processing, and specifically to a cross-source data collaborative analysis method and system based on an AI model. Background Art

[0002] With the advent of the big data era, data has become an important asset for enterprises and society. However, data is usually scattered in different data sources, such as different databases, file systems, cloud platforms, etc. There are problems such as inconsistent data formats, uneven data quality, and limited data access rights among these data sources, making cross-source data collaborative analysis very difficult. In recent years, the development of artificial intelligence technology has provided new ideas and methods for cross-source data collaborative analysis.

[0003] For example, the invention patent with announcement number: CN116910161B is a collaborative analysis system, method, electronic device and computer-readable medium, including a capability management center, a collaborative service center and several data centers, each data center includes a modeling platform, a model collaborative engine, a model service, and a data service, wherein the model service includes an online modeling engine center, an offline modeling engine center and a real-time modeling engine center; the collaborative analysis system includes an upper and lower cascade mode and a cross-network collaborative mode. In the upper and lower cascade mode, several data centers share the same capability management center and collaborative service center, and perform call management or collaborative service release with each other; in the cross-network collaborative mode, the collaborative analysis system has a security boundary, and the data centers on both sides of the security boundary perform call management or collaborative service release with the capability management center and collaborative service center on the corresponding side, and interact through the capability management center and collaborative service center on each side.

[0004] For example, the invention patent with announcement number: CN113886457B is a method for solving cross-domain heterogeneous data joint retrieval, which has the following steps: S1, build an engine management platform; S2, build a distributed data collaboration engine; S3, implement a data access engine; S4, a multi-domain data source management platform; S5, build a collaborative data cache library; S6, open a collaborative service bus; S7, output a complete collaborative computing bus platform.

[0005] However, it was found that the above technology has at least the following technical problems: the existing cross-source data collaborative analysis methods based on AI models have problems with insufficient data sources and insufficient data real-time performance in data collection, which often have a negative impact on model application and user experience. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a cross-source data collaborative analysis method and system based on an AI model, which can effectively solve the problems involved in the above-mentioned background technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: The first aspect of the present invention provides a cross-source data collaborative analysis method based on an AI model, comprising: real-time tracking of emerging smart home product sales channels, and regular value assessment of each sales channel, screening out platforms with data collection value, marking them as newly added platforms, and jointly marking them with the original data collection platforms as data collection platforms; Develop a real-time monitoring algorithm for data changes, analyze the changing trends of smart home product retail data on each data collection platform, automatically adjust the data collection frequency for data collection, and obtain multi-source data; Use a distributed network architecture for data transmission, monitor data transmission status information in real time, determine whether data transmission is abnormal, and provide early warning feedback.

[0008] As a further method, the value assessment of each sales channel is performed regularly, and the specific process is: regularly collecting sales data reports and user information of each sales channel according to a preset time period, performing comprehensive processing to obtain the value demand assessment index of each sales channel, and performing value assessment on each sales channel according to the value demand assessment index of each sales channel; The sales data report includes: sales amount, sales volume, and sales proportion of different products; The user information includes: user browsing volume, transaction volume, satisfaction score, and the ratio of review users to total purchasing users; Sort the value demand assessment indicators of each original data collection platform from large to small, obtain the lowest value of the value demand assessment indicator sequence, and mark it as the value demand assessment indicator threshold; The value demand assessment indicators of each sales channel are compared with the value demand assessment indicator threshold. If the value demand assessment indicator of a sales channel is higher than or equal to the value demand assessment indicator threshold, the value assessment of the sales channel is qualified; if the value demand assessment indicator of a sales channel is lower than the value demand assessment indicator threshold, the value assessment of the sales channel is unqualified.

[0009] As a further method, the screening obtains platforms with data collection value, and the specific analysis process is: statistically evaluate the sales channels as qualified, mark them as qualified sales channels, obtain the resource input of each qualified sales channel, and compare it with the average of the resource input of the original data collection platform. If the resource input of a qualified sales channel is lower than or equal to the average of the resource input of the original data collection platform, then the qualified sales channel is marked as a new platform and data is collected for it. If the resource input of a qualified sales channel is higher than the average of the resource input of the original data collection platform, then data is not collected for it. Each newly added platform and the original data collection platform are jointly labeled as data collection platforms.

[0010] As a further method, the change trend of the smart home product retail data of each data collection platform is analyzed, and the specific analysis process is: obtaining the historical data of the smart home product retail of each data collection platform, and analyzing the fluctuation degree of the historical data of each data collection platform; A real-time monitoring algorithm for data changes is developed by combining time series analysis algorithm and machine learning algorithm. The algorithm is used to predict the fluctuation degree of smart home product retail data of each data collection platform in the future, and the prediction results are obtained and marked as the predicted data fluctuation degree of each data collection platform.

[0011] As a further method, the data collection frequency is automatically adjusted to collect data, and the specific analysis process is: according to the historical data fluctuation degree of each data collection platform, the initial data collection frequency is set; Adjust the frequency of initial data collection based on the degree of fluctuation of the predicted data of each data collection platform; If the volatility of the predicted data is greater than that of the historical data, the frequency of data collection will be increased; if the volatility of the predicted data is less than that of the historical data, the frequency of data collection will be reduced; Establish a feedback mechanism to compare the actual data fluctuation with the predicted data fluctuation and continuously optimize the data collection frequency; If the actual data fluctuation degree is greater than the predicted data fluctuation degree, it means that the current data collection frequency cannot meet the demand, and the data collection frequency should be increased in time. Conversely, if the actual data fluctuation degree is less than the predicted data fluctuation degree, the data collection frequency should be reduced.

[0012] As a further method, the real-time monitoring of data transmission status information determines whether the data transmission is abnormal and provides early warning feedback, wherein the data transmission status information includes the sales data transmission rate at each monitoring time point, the sales data packet loss rate at each monitoring time point and the original received data of each data collection platform; Obtaining preset sales data transmission rate threshold, sales data packet loss rate threshold and primary verification value of each data collection platform from the database; The sales data transmission rate at each monitoring time point is compared with the sales data transmission rate threshold. If the sales data transmission rate at a certain monitoring time point is greater than or equal to the sales data transmission rate threshold, no additional operation is performed. If the sales data transmission rate at a certain monitoring time point is less than the sales data transmission rate threshold, an abnormal data transmission delay warning is issued. The sales data packet loss rate at each monitoring time point is compared with the sales data packet loss rate threshold. If the sales data packet loss rate at a certain monitoring time point is greater than or equal to the sales data packet loss rate threshold, no additional operation is performed. If the sales data packet loss rate at a certain monitoring time point is less than the sales data packet loss rate threshold, an abnormal data packet loss warning is issued. The sales data receiving end obtains the original received data of each data collection platform, and performs data separation on the original received data of each data collection platform to obtain the original data and verification value of the original received data of each data collection platform. At the same time, the sales data receiving end uses a preset verification algorithm to process the original data of the original received data of each data collection platform to obtain the secondary verification value of each data collection platform; The secondary verification value of each data collection platform is compared with the primary verification value of each data collection platform. If the two verification values ​​are exactly the same, no additional operation is performed. If the comparison results are inconsistent, a transmission data verification error abnormality warning is issued.

[0013] As a further method, the data transmission delay abnormality warning, the specific analysis process is: obtain data transmission delay abnormality information, where the data transmission delay abnormality information includes the platform name of the data transmission delay abnormality, the delay time and the potential affected business scope, package to obtain the data transmission delay abnormality data set, and transmit the data transmission delay abnormality data set to the mobile terminal of the data collection and transmission management team, and at the same time identify the abnormal platform on the data monitoring system interface, and display the data transmission delay abnormality information.

[0014] As a further method, the packet loss anomaly warning, the specific analysis process is: obtain packet loss anomaly information, where the packet loss anomaly information includes the platform of the packet loss anomaly, the loss rate value and the potential impact on the business scope, generate an anomaly report in the data monitoring system, record the time of the anomaly, duration and number of lost packets, and send warning information to technical personnel, the technical personnel check the network equipment status and transmission link according to the packet loss anomaly information.

[0015] As a further method, the transmission data verification error abnormal warning, the specific analysis process is: obtaining transmission data verification error abnormal information, wherein the transmission data verification error abnormal information includes the platform and data batch where the verification error occurs, sending an alarm to the data security management team, and suspending data collection from the platform; The data security management team backs up abnormal data and performs data checks based on abnormal information of transmission data verification errors.

[0016] The second aspect of the present invention provides a cross-source data collaborative analysis system based on an AI model, comprising: a data source optimization module, a data acquisition optimization module, and a data transmission optimization module; The data source optimization module is used to track the sales channels of emerging smart home products in real time, and regularly evaluate the value of each sales channel, screen out platforms with data collection value, mark them as new platforms, and jointly mark them with the original data collection platforms as data collection platforms; The data collection optimization module is used to develop a real-time monitoring algorithm for data changes, analyze the changing trends of smart home product retail data on each data collection platform, automatically adjust the data collection frequency for data collection, and obtain multi-source data; The data transmission optimization module is used to adopt a distributed network architecture for data transmission, monitor the data transmission status information in real time, determine whether the data transmission is abnormal, and provide early warning feedback.

[0017] Compared with the prior art, the present invention has at least the following advantages or beneficial effects: (1) The present invention provides a cross-source data collaborative analysis method and system based on an AI model, tracks and evaluates the value of sales channels in real time, screens new data collection platforms, optimizes data sources, helps improve data representativeness and quality, develops algorithms to automatically adjust data collection frequency, obtains multi-source data, improves data richness, timeliness and quality, uses a distributed network architecture to transmit data, monitors and warns of anomalies in real time, and ensures stable and accurate data transmission; (2) The present invention collects smart home product retail data from various data collection platforms through multiple sources, and automatically adjusts the data collection frequency according to data fluctuations, which is conducive to improving data richness, timeliness and quality, helping AI models adapt to complex markets, optimizing training and performance, and mining data associations, thereby providing more comprehensive and accurate information for AI models to achieve cross-source data collaborative analysis, and enhancing analysis effects and efficiency; (3) The present invention determines whether the transmission is abnormal and issues an early warning by comparing the sales data transmission rate, packet loss rate and check value. In the event of an abnormality, the present invention feeds back detailed abnormal information to the corresponding team, performs operations such as identification, inspection, suspension of collection and backup, thereby ensuring stable and accurate data transmission, reducing the impact of abnormalities on the AI ​​model, providing it with high-quality multi-source data, helping the AI ​​model to mine data associations, and achieving more effective cross-source data collaborative analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0019] Figure 1 It is a schematic diagram of the method flow of the present invention.

[0020] Figure 2 It is a schematic diagram of system module connection of the present invention. DETAILED DESCRIPTION

[0021] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0022] Reference Figure 1 As shown, the first aspect of the present invention provides a cross-source data collaborative analysis method based on an AI model, which tracks the sales channels of emerging smart home products in real time, and regularly evaluates the value of each sales channel, screens out platforms with data collection value, marks them as newly added platforms, and jointly marks them with the original data collection platforms as data collection platforms.

[0023] It needs to be explained that the emerging smart home product sales channels specifically refer to major e-commerce platforms.

[0024] Specifically, a value assessment is conducted on each sales channel on a regular basis. The specific process is: sales data reports and user information of each sales channel are regularly collected according to a preset time period, and comprehensive processing is performed to obtain a value demand assessment index for each sales channel; and a value assessment is conducted on each sales channel based on the value demand assessment index for each sales channel.

[0025] The sales data report includes: sales revenue, sales volume, and sales proportion of different products.

[0026] The user information includes: user browsing volume, transaction volume, satisfaction score, and the ratio of evaluating users to total purchasing users.

[0027] Sort the value demand assessment indicators of each original data collection platform from large to small, obtain the lowest value of the value demand assessment indicator sequence, and mark it as the value demand assessment indicator threshold.

[0028] The value demand assessment indicators of each sales channel are compared with the value demand assessment indicator threshold. If the value demand assessment indicator of a sales channel is higher than or equal to the value demand assessment indicator threshold, the value assessment of this sales channel is qualified; if the value demand assessment indicator of a sales channel is lower than the value demand assessment indicator threshold, the value assessment of this sales channel is unqualified.

[0029] Furthermore, platforms with data collection value are screened out, and the specific analysis process is as follows: sales channels that are evaluated as qualified by statistical value are marked as qualified sales channels, the resource input of each qualified sales channel is obtained, and compared with the mean resource input of the original data collection platform. If the resource input of a qualified sales channel is lower than or equal to the mean resource input of the original data collection platform, the qualified sales channel is marked as a new platform and data collection is performed on it; if the resource input of a qualified sales channel is higher than the mean resource input of the original data collection platform, data collection is not performed on it.

[0030] Each newly added platform and the original data collection platform are jointly labeled as data collection platforms.

[0031] It is necessary to explain that by real-time tracking of the sales of emerging smart home products on major e-commerce platforms, i.e., emerging smart home product sales channels, it is possible to timely understand the sales performance of products on different platforms, thereby more accurately grasping market dynamics and gaining insight into changes in consumer demand and market trends. For example, if the sales volume of smart home products on a certain emerging platform suddenly increases, it may mean that the user group of the platform has a high interest in and demand for such products, and companies can adjust their market strategies accordingly.

[0032] It should be explained that by regularly conducting value assessments on various sales channels and selecting platforms with data collection value based on value demand assessment indicators, the allocation of data collection resources can be effectively optimized. Resources are concentrated on platforms with high value and relatively reasonable resource investment, avoiding wasting resources on platforms with low value or excessive resource investment. For example, by comparing the value demand assessment indicators and resource investment of each platform, newly added platforms are marked for data collection, which improves the efficiency and pertinence of data collection and makes the collected data more representative and valuable for analysis. Furthermore, by comprehensively considering sales data reports (sales, sales volume, sales share of different products) and user information (user views, transaction volume, satisfaction rating, the ratio of evaluated users to total purchasing users) for value assessment, the value of sales channels can be comprehensively measured from multiple dimensions. The data of the data collection platform selected in this way can better reflect the real situation of the market and the real needs of users, thereby improving the quality of the data and providing a more reliable basis for subsequent data analysis and decision-making.

[0033] It needs to be explained that with the development and changes of the market, new e-commerce platforms continue to emerge. Through this real-time tracking and regular evaluation mechanism, new platforms with potential can be discovered in a timely manner and incorporated into the data collection platform system, so that enterprises can better adapt to market changes, maintain a keen insight into the market, adjust business strategies in a timely manner, and enhance market competitiveness.

[0034] Develop a real-time monitoring algorithm for data changes, analyze the changing trends of smart home product retail data on each data collection platform, automatically adjust the data collection frequency for data collection, and obtain multi-source data.

[0035] Specifically, the change trend of the smart home product retail data of each data collection platform is analyzed. The specific analysis process is: obtaining the historical data of the smart home product retail of each data collection platform, and analyzing the fluctuation degree of the historical data of each data collection platform; A real-time monitoring algorithm for data changes is developed by combining time series analysis algorithm and machine learning algorithm. The algorithm is used to predict the fluctuation degree of smart home product retail data of each data collection platform in the future, and the prediction results are obtained and marked as the predicted data fluctuation degree of each data collection platform.

[0036] Furthermore, the data collection frequency is automatically adjusted for data collection. The specific analysis process is as follows: according to the historical data fluctuation degree of each data collection platform, the initial data collection frequency is set; Adjust the frequency of initial data collection based on the degree of fluctuation of the predicted data of each data collection platform; If the volatility of the predicted data is greater than that of the historical data, the frequency of data collection will be automatically increased; if the volatility of the predicted data is less than that of the historical data, the frequency of data collection will be reduced; Establish a feedback mechanism to compare the actual data fluctuation with the predicted data fluctuation and continuously optimize the data collection frequency; If the actual data fluctuation degree is greater than the predicted data fluctuation degree, it means that the current data collection frequency may not meet the demand, and the data collection frequency should be increased in time. Conversely, if the actual data fluctuation degree is less than the predicted data fluctuation degree, the data collection frequency should be appropriately reduced to save resources.

[0037] In a specific embodiment, in the early stage of product sales, due to the uncertainty of market response and large data fluctuations, a higher data collection frequency is set, such as once every hour, so as to capture the dynamic changes of the market in a timely manner and provide sufficient data for subsequent analysis.

[0038] It should be explained that in this embodiment, by collecting smart home product retail data from various data collection platforms from multiple sources, richer and more comprehensive data can be provided for the AI ​​model. The data from different platforms contain information such as different user groups, sales models, market reactions, etc. The integration of these multi-source data enables the AI ​​model to understand the market situation of smart home products from multiple dimensions, providing more sufficient materials for cross-source data collaborative analysis, and helping the model to dig out deeper and more valuable information and rules. At the same time, the data collection frequency is automatically adjusted according to the degree of data fluctuation, and the collection frequency is increased when the data fluctuation is large, ensuring the timeliness of the data. Timely data is crucial for the cross-source data collaborative analysis of the AI ​​model, which enables the model to capture market changes and the dynamic trends of data from each platform more quickly, making the analysis results closer to the actual situation and improving the model's response speed and adaptability to market changes.

[0039] Furthermore, by setting the initial data collection frequency based on the historical data fluctuations, and then adjusting the data collection frequency based on the predicted and actual data fluctuations, the problem of insufficient or excessive data collection can be effectively avoided. This helps improve the quality of data, reduce data noise and bias, and provide a more reliable data foundation for cross-source data collaborative analysis of AI models. High-quality data can make model training and prediction more accurate, improve model performance and the credibility of analysis results.

[0040] It needs to be explained that accurate and timely data can enable AI models to learn more accurate patterns and rules during the training process, avoiding model bias caused by incomplete or inaccurate data. By continuously optimizing the frequency of data collection and providing high-quality multi-source data, it helps AI models to better integrate and analyze cross-source data, improve the generalization ability and prediction accuracy of the model, and thus improve the effect and efficiency of cross-source data collaborative analysis. More importantly, the collection and integration of multi-source data provides AI models with more data dimensions, which helps the model to explore potential correlations and synergies between data from different platforms. By analyzing the changing trends and relationships of data from each platform, AI models can discover the complementarity and synergy opportunities of different platforms in the sales of smart home products, provide valuable references for companies to optimize sales channels and integrate resources, and achieve more efficient cross-source data collaborative analysis.

[0041] Use a distributed network architecture for data transmission, monitor data transmission status information in real time, determine whether data transmission is abnormal, and provide early warning feedback.

[0042] Specifically, the data transmission status information is monitored in real time to determine whether the data transmission is abnormal and to provide early warning feedback, wherein the data transmission status information includes the sales data transmission rate at each monitoring time point, the sales data packet loss rate at each monitoring time point and the original received data of each data collection platform.

[0043] The preset sales data transmission rate threshold, sales data packet loss rate threshold and the first-level verification value of each data collection platform are obtained from the database.

[0044] The sales data transmission rate at each monitoring time point is compared with the sales data transmission rate threshold. If the sales data transmission rate at a certain monitoring time point is greater than or equal to the sales data transmission rate threshold, no additional operation is performed; if the sales data transmission rate at a certain monitoring time point is less than the sales data transmission rate threshold, an abnormal data transmission delay warning is issued.

[0045] The sales data packet loss rate at each monitoring time point is compared with the sales data packet loss rate threshold. If the sales data packet loss rate at a certain monitoring time point is greater than or equal to the sales data packet loss rate threshold, no additional operation is performed; if the sales data packet loss rate at a certain monitoring time point is less than the sales data packet loss rate threshold, an abnormal data packet loss warning is issued.

[0046] The sales data receiving end obtains the original received data of each data collection platform, and separates the original received data of each data collection platform to obtain the original data and verification value of the original received data of each data collection platform. At the same time, the sales data receiving end uses a preset verification algorithm to process the original data of the original received data of each data collection platform to obtain the secondary verification value of each data collection platform.

[0047] The secondary verification value of each data collection platform is compared with the primary verification value of each data collection platform. If the two verification values ​​are exactly the same, no additional operation is performed; if the comparison results are inconsistent, a transmission data verification error abnormality warning is issued.

[0048] Furthermore, the data transmission delay anomaly warning, the specific analysis process is: obtain the data transmission delay anomaly information, where the data transmission delay anomaly information includes the platform name of the data transmission delay anomaly, the delay time and the potential impact on the business scope, package to obtain the data transmission delay anomaly data set, and transmit the data transmission delay anomaly data set to the mobile terminal of the data collection and transmission management team, and at the same time identify the abnormal platform on the data monitoring system interface, and display the data transmission delay anomaly information.

[0049] Furthermore, the data packet loss anomaly warning, the specific analysis process is: obtain the data packet loss anomaly information, where the data packet loss anomaly information includes the platform of the data packet loss anomaly, the loss rate value and the potential impact on the business scope, generate an anomaly report in the data monitoring system, record the time of the anomaly, duration and number of lost data packets, and send warning information to technical personnel, who check the network equipment status and transmission link based on the data packet loss anomaly information.

[0050] Specifically, the transmission data verification error abnormal warning, the specific analysis process is: obtain the transmission data verification error abnormal information, where the transmission data verification error abnormal information includes the platform and data batch where the verification error occurs, send an alarm to the data security management team, and suspend data collection from the platform.

[0051] The data security management team backs up abnormal data and performs data checks based on abnormal information of transmission data verification errors.

[0052] It should be explained that this embodiment adopts a distributed network architecture for data transmission, and through the monitoring and early warning of sales data transmission rate, packet loss rate and data verification, it is conducive to ensuring the stability and accuracy of data transmission, avoiding incomplete and inaccurate data due to transmission delays, packet loss or verification errors, and is conducive to ensuring the consistency and reliability of cross-source data. By formulating a detailed early warning feedback mechanism for data transmission anomalies, such as sending abnormal information to relevant teams and generating reports, problems can be discovered and handled in a timely manner, helping to reduce the impact of data transmission problems on AI model analysis and improve the efficiency of model training and analysis. At the same time, by suspending data collection on the platform where verification errors occur and backing up abnormal data, it helps the data security management team to conduct targeted inspections and maintain data quality, thereby helping the AI ​​model to mine the associations and laws between different source data based on accurate data, achieve more effective cross-source data collaborative analysis, and improve the credibility and practicality of model analysis results.

[0053] Reference Figure 2 As shown, the second aspect of the present invention provides a cross-source data collaborative analysis system based on an AI model, comprising: a data source optimization module, a data acquisition optimization module, a data transmission optimization module and a database.

[0054] The data source optimization module is used to track the sales channels of emerging smart home products in real time, and regularly evaluate the value of each sales channel, screen out platforms with data collection value, mark them as new platforms, and jointly mark them with the original data collection platforms as data collection platforms.

[0055] The data collection optimization module is used to develop a real-time monitoring algorithm for data changes, analyze the changing trends of smart home product retail data on each data collection platform, automatically adjust the data collection frequency for data collection, and obtain multi-source data.

[0056] The data transmission optimization module is used to adopt a distributed network architecture for data transmission, monitor the data transmission status information in real time, determine whether the data transmission is abnormal, and provide early warning feedback.

[0057] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.

Claims

1. A cross-source data collaborative analysis method based on an AI model, characterized in that: include: Track emerging smart home product sales channels in real time, conduct value assessments on each sales channel on a regular basis, screen out platforms with data collection value, mark them as new platforms, and jointly mark them with the original data collection platforms as data collection platforms; Develop a real-time monitoring algorithm for data changes, analyze the changing trends of smart home product retail data on each data collection platform, automatically adjust the data collection frequency for data collection, and obtain multi-source data; Use a distributed network architecture for data transmission, monitor data transmission status information in real time, determine whether data transmission is abnormal, and provide early warning feedback.

2. The cross-source data collaborative analysis method based on the AI ​​model according to claim 1 is characterized by: The specific process of regularly evaluating the value of each sales channel is as follows: Collect sales data reports and user information of each sales channel regularly according to the preset time period, conduct comprehensive processing to obtain the value demand evaluation index of each sales channel, and conduct value evaluation of each sales channel according to the value demand evaluation index of each sales channel; The sales data report includes: sales amount, sales volume, and sales proportion of different products; The user information includes: user browsing volume, transaction volume, satisfaction score, and the ratio of review users to total purchasing users; Sort the value demand assessment indicators of each original data collection platform from large to small, obtain the lowest value of the value demand assessment indicator sequence, and mark it as the value demand assessment indicator threshold; The value demand assessment indicators of each sales channel are compared with the value demand assessment indicator threshold. If the value demand assessment indicator of a sales channel is higher than or equal to the value demand assessment indicator threshold, the value assessment of this sales channel is qualified; if the value demand assessment indicator of a sales channel is lower than the value demand assessment indicator threshold, the value assessment of this sales channel is unqualified.

3. The cross-source data collaborative analysis method based on the AI ​​model according to claim 1 is characterized in that: The screening process obtains a platform with data collection value, and the specific analysis process is as follows: The sales channels that are evaluated as qualified by statistical value are marked as qualified sales channels, and the resource input of each qualified sales channel is obtained and compared with the average resource input of the original data collection platform. If the resource input of a qualified sales channel is lower than or equal to the average resource input of the original data collection platform, the qualified sales channel is marked as a new platform and data is collected for it; if the resource input of a qualified sales channel is higher than the average resource input of the original data collection platform, data is not collected for it; Each newly added platform and the original data collection platform are jointly labeled as data collection platforms.

4. The cross-source data collaborative analysis method based on the AI ​​model according to claim 1 is characterized in that: The change trend of the smart home product retail data of each data collection platform is analyzed, and the specific analysis process is as follows: Obtain the historical data of smart home product retail sales from each data collection platform, and analyze the fluctuation degree of the historical data from each data collection platform; A real-time monitoring algorithm for data changes is developed by combining time series analysis algorithm and machine learning algorithm. The algorithm is used to predict the fluctuation degree of smart home product retail data of each data collection platform in the future, and the prediction results are obtained and marked as the predicted data fluctuation degree of each data collection platform.

5. The cross-source data collaborative analysis method based on the AI ​​model according to claim 1 is characterized in that: The automatic adjustment of data collection frequency for data collection, the specific analysis process is as follows: Set the initial data collection frequency based on the historical data fluctuation level of each data collection platform; Adjust the frequency of initial data collection based on the degree of fluctuation of the predicted data of each data collection platform; If the volatility of the predicted data is greater than that of the historical data, the frequency of data collection will be increased; if the volatility of the predicted data is less than that of the historical data, the frequency of data collection will be reduced; Establish a feedback mechanism to compare the actual data fluctuation with the predicted data fluctuation and continuously optimize the data collection frequency; If the actual data fluctuation degree is greater than the predicted data fluctuation degree, it means that the current data collection frequency cannot meet the demand, and the data collection frequency should be increased in time. Conversely, if the actual data fluctuation degree is less than the predicted data fluctuation degree, the data collection frequency should be reduced.

6. The cross-source data collaborative analysis method based on the AI ​​model according to claim 1 is characterized in that: The real-time monitoring of data transmission status information determines whether the data transmission is abnormal and provides early warning feedback, wherein the data transmission status information includes the sales data transmission rate at each monitoring time point, the sales data packet loss rate at each monitoring time point and the original received data of each data collection platform; Obtaining preset sales data transmission rate threshold, sales data packet loss rate threshold and primary verification value of each data collection platform from the database; The sales data transmission rate at each monitoring time point is compared with the sales data transmission rate threshold. If the sales data transmission rate at a certain monitoring time point is greater than or equal to the sales data transmission rate threshold, no additional operation is performed; if the sales data transmission rate at a certain monitoring time point is less than the sales data transmission rate threshold, an abnormal data transmission delay warning is issued; The sales data packet loss rate at each monitoring time point is compared with the sales data packet loss rate threshold. If the sales data packet loss rate at a certain monitoring time point is greater than or equal to the sales data packet loss rate threshold, no additional operation is performed; if the sales data packet loss rate at a certain monitoring time point is less than the sales data packet loss rate threshold, an abnormal data packet loss warning is issued; The sales data receiving end obtains the original received data of each data collection platform, and performs data separation on the original received data of each data collection platform to obtain the original data and verification value of the original received data of each data collection platform. At the same time, the sales data receiving end uses a preset verification algorithm to process the original data of the original received data of each data collection platform to obtain the secondary verification value of each data collection platform; The secondary verification value of each data collection platform is compared with the primary verification value of each data collection platform. If the two verification values ​​are exactly the same, no additional operation is performed; if the comparison results are inconsistent, a transmission data verification error abnormality warning is issued.

7. The cross-source data collaborative analysis method based on the AI ​​model according to claim 6 is characterized by: The specific analysis process of the abnormal data transmission delay warning is as follows: Obtain data transmission delay anomaly information, including the platform name, delay time, and potentially affected business scope of the data transmission delay anomaly, package it to obtain a data transmission delay anomaly data set, and transmit the data transmission delay anomaly data set to the mobile terminal of the data collection and transmission management team. At the same time, identify the abnormal platform on the data monitoring system interface and display the data transmission delay anomaly information.

8. The cross-source data collaborative analysis method based on the AI ​​model according to claim 6 is characterized by: The specific analysis process of the abnormal warning of data packet loss is as follows: Obtain data packet loss exception information, including the platform of data packet loss exception, loss rate value and potential impact on business scope, generate an exception report in the data monitoring system, record the time, duration and number of lost data packets, and send early warning information to technical personnel. Technical personnel check the network equipment status and transmission link according to the data packet loss exception information.

9. The cross-source data collaborative analysis method based on the AI ​​model according to claim 6 is characterized by: The specific analysis process of the transmission data verification error abnormal warning is as follows: Obtaining transmission data verification error exception information, where the transmission data verification error exception information includes the platform and data batch where the verification error occurs, sending an alarm to the data security management team, and suspending data collection from the platform; The data security management team backs up abnormal data and performs data checks based on abnormal information of transmission data verification errors.

10. A cross-source data collaborative analysis system based on an AI model, applying the cross-source data collaborative analysis method based on an AI model as described in any one of claims 1 to 9, characterized in that: include: The data source optimization module is used to track the sales channels of emerging smart home products in real time, and regularly evaluate the value of each sales channel, screen out platforms with data collection value, mark them as new platforms, and jointly mark them with the original data collection platforms as data collection platforms; The data collection optimization module is used to develop a real-time monitoring algorithm for data changes, analyze the changing trends of smart home product retail data on each data collection platform, automatically adjust the data collection frequency for data collection, and obtain multi-source data; The data transmission optimization module is used to adopt a distributed network architecture for data transmission, monitor the data transmission status information in real time, determine whether the data transmission is abnormal, and provide early warning feedback.

Images