Data processing method and system based on intelligent algorithm

Abstract

The invention discloses a data processing method and system based on an intelligent algorithm. The method comprises the following steps: collecting transaction data in real time; automatically identifying an application scene and a transaction object of the preprocessed transaction data by utilizing a machine learning technology; preprocessing the collected transaction data by using an edge computing device with a built-in lightweight AI model to obtain preprocessed transaction data; according to the identified application scene and transaction object of the transaction data, selecting an intelligent algorithm model matched with the identified application scene and transaction object of the transaction data from an intelligent algorithm library, and analyzing the preprocessed transaction data by using the matched intelligent algorithm model to obtain a transaction data analysis result; and respectively storing the transaction data analysis results obtained based on different transaction objects, and visually displaying the transaction data analysis results to the corresponding transaction terminals. According to the invention, efficient and high-adaptability data processing can be realized.

Description

Technical Field

[0001] The present application relates to the technical field of big data processing, and specifically to a data processing method and system based on intelligent algorithms. Background Art

[0002] With the rapid development of Internet technology, transaction data processing has become one of the key links in the security and efficiency of trading systems. Effective transaction data processing methods can not only improve the transparency of the trading market, but also help trading institutions to discover potential risks in a timely manner, thereby improving the overall transaction security. Therefore, how to efficiently and accurately process massive transaction data has become one of the problems that need to be solved in various industries.

[0003] Existing transaction data processing technologies mainly rely on traditional data mining methods, such as centrally storing historical transaction data in data warehouses, and then using batch processing to clean, integrate and analyze data. In addition, some applications will combine statistical methods to pre-process transaction data and use machine learning algorithms to perform preliminary data analysis and prediction. These methods have improved the speed and accuracy of data processing to a certain extent, but there are still some shortcomings.

[0004] The above traditional methods often fail to achieve an effective combination of real-time and accuracy when processing large-scale transaction data with high real-time requirements. Delays are prone to occur in the data preprocessing stage. When faced with dynamically changing application scenarios and different transaction objects, their adaptability is poor and it is difficult to quickly adjust strategies to cope with new transaction data types and scenario requirements. Summary of the invention

[0005] In order to achieve efficient and adaptable data processing, the present application provides a data processing method and system based on an intelligent algorithm.

[0006] In a first aspect, the present application provides a data processing method based on an intelligent algorithm, comprising: Collect transaction data in real time; Use machine learning technology to automatically identify the application scenarios and transaction objects of pre-processed transaction data; Use edge computing devices with built-in lightweight AI models to pre-process the collected transaction data and obtain pre-processed transaction data; According to the application scenarios and transaction objects of the identified transaction data, an intelligent algorithm model matching the application scenarios and transaction objects of the identified transaction data is selected from the intelligent algorithm library, and the pre-processed transaction data is analyzed using the matching intelligent algorithm model to obtain transaction data analysis results; the intelligent algorithm library stores a number of intelligent algorithms generated by training using training data containing historical transaction data analysis results of different application scenarios and transaction objects, as well as a mapping table of historical application scenarios and transaction objects and intelligent algorithms; The transaction data analysis results obtained based on different transaction objects are stored separately and displayed visually to the corresponding transaction terminals.

[0007] By adopting the above solution, edge computing devices with built-in lightweight AI models are used to pre-process the collected transaction data, shorten the data processing cycle, and improve the real-time performance of information processing; by selecting intelligent algorithm models that match the application scenarios and transaction objects of the identified transaction data from the intelligent algorithm library, the optimal intelligent algorithm model is dynamically selected for analysis according to different application scenarios and transaction objects, thereby improving the accuracy and reliability of the analysis results; storage and visualization are performed based on the adaptability of different transaction objects, It is convenient for corresponding users to intuitively understand their own related transaction status.

[0008] Preferably, the real-time collection of transaction data specifically includes: The knowledge topology map of each trading platform is obtained, and the correlation score of each trading platform in the knowledge topology map is obtained by using the graph analysis method, and then each trading platform with a correlation score greater than the preset score is obtained as a strongly correlated trading platform in the knowledge topology map; the API interface of each trading platform is used to complete the acquisition of cross-platform transaction data of strongly correlated trading platforms, and the collected transaction data is converted and exchanged in accordance with the standardized data format in the cross-platform transaction data sharing agreement.

[0009] By adopting the above solution, the cross-platform transaction data of strongly related trading platforms is effectively collected and standardized using the trading platform knowledge topology map, thereby improving the accuracy and efficiency of transaction data collection.

[0010] Preferably, the edge computing device with a built-in lightweight AI model performs abnormal data cleaning and preprocessing on the collected transaction data, specifically including: The edge computing device with a built-in lightweight AI model may adopt different AI models, including some AI models generated by training transaction data in a fixed application scenario after preprocessing; determining the edge computing device with the same application scenario as the transaction data in each edge device according to the application scenario of the identified transaction data; counting the number of collected transaction data, and determining whether the computing resources of the determined edge computing device meet the preprocessing of the current transaction data volume. If not, using the remaining AI models for preprocessing.

[0011] By adopting the above solution, the appropriate AI model is selected for preprocessing according to the application scenario of the transaction data to ensure that the preprocessing process is efficient and accurate; at the same time, when the computing resources of the determined edge computing device are insufficient, it can automatically switch to other AI models, ensuring the stability of the preprocessing process and the integrity of the transaction data analysis.

[0012] Preferably, the analyzing the pre-processed transaction data using the matching intelligent algorithm model further includes: The features of the preprocessed transaction data, including data volume, data distribution and speed, are extracted using a matched intelligent algorithm model; the corresponding intelligent algorithm model parameters are matched according to the features of the extracted preprocessed transaction data to complete the parameter adjustment of the intelligent algorithm model, and the analysis of the preprocessed transaction data is continued with the intelligent algorithm model after parameter adjustment; the intelligent algorithm model parameters matched with the features of the extracted preprocessed transaction data are the intelligent model parameters with the highest user satisfaction score in the analysis results obtained by the intelligent algorithm model corresponding to the intelligent algorithm model parameters matched according to the historical matching.

[0013] By adopting the above solution, the corresponding intelligent algorithm model parameters are matched according to the extracted features, and the parameter adjustment of the intelligent algorithm model is completed, thereby improving the accuracy and efficiency of the intelligent algorithm model in analyzing the preprocessed transaction data; based on the highest satisfaction score feedback from users in the analysis results corresponding to the historical intelligent algorithm model parameters, the effectiveness and reliability of the intelligent algorithm model parameter adjustment are further guaranteed.

[0014] Preferably, it also includes: The intelligent algorithm model selected from the intelligent algorithm library to match the application scenario and transaction object of the identified transaction data includes an intelligent algorithm model matching the application scenario of the identified transaction data and a unilateral transaction object, and an intelligent algorithm model matching the application scenario of the identified transaction data and a multi-party transaction object; When a timeout or failure occurs when analyzing the pre-processed transaction data using an intelligent algorithm model that matches the application scenario of the identified transaction data and multi-party transaction objects, check whether an intelligent algorithm model that matches the application scenario of the identified transaction data and a unilateral transaction object is also matched and whether a timeout or failure occurs when analyzing the data using the intelligent algorithm model that matches the application scenario of the identified transaction data and the unilateral transaction object. If a timeout or failure occurs again, it is determined that there is a problem with the corresponding unilateral transaction object data, and the transaction data is preprocessed again. Otherwise, a spare intelligent algorithm model that matches the application scenario of the identified transaction data and multi-party transaction objects is directly selected to complete the analysis of the transaction data.

[0015] By adopting the above solution, the transaction data analysis results are compared between the intelligent algorithm model that matches the application scenario of the identified transaction data and multi-party transaction objects and the intelligent algorithm model that matches the application scenario of the identified transaction data and a single-party transaction object. It is possible to determine whether the overall analysis failure is caused by the abnormal data of a single transaction object or whether there is a problem with the model itself and deal with it in a targeted manner, thereby improving the accuracy and robustness of the transaction data analysis.

[0016] Preferably, it also includes: Obtain the transaction data analysis results of the specified transaction objects other than itself by the transaction object, and determine whether the current transaction object has the preset activation authority to obtain the transaction data analysis results of the specified transaction object. If so, the transaction data analysis results of the specified transaction object are returned to the transaction object terminal that sent the acquisition request.

[0017] By adopting the above solution, taking into account the needs of different transaction objects to obtain transaction data analysis results of designated transaction objects other than themselves, the transaction data analysis results can be displayed according to user needs based on the approval of authority, thereby realizing controllable sharing of data between transaction objects and improving data security and privacy protection.

[0018] Preferably, it also includes: Obtain the transaction object's request to share its own transaction data analysis results, determine whether the current transaction object has the preset activation authority to share its own transaction data analysis results, and if so, transmit the transaction object's own transaction data analysis results to all transaction object terminals for display.

[0019] By adopting the above solution, taking into account the need of some transaction parties to share data analysis results, while enhancing data transparency and collaboration efficiency, shared audits are also conducted to avoid illegal disclosure of data analysis results and ensure data security.

[0020] In a second aspect, the present application provides a data processing system based on an intelligent algorithm, comprising: Transaction data collection module, used to collect transaction data in real time; The transaction data classification module is used to automatically identify the application scenarios and transaction objects of the pre-processed transaction data using machine learning technology; A transaction data processing module is used to pre-process the collected transaction data using an edge computing device with a built-in lightweight AI model to obtain the pre-processed transaction data; A transaction data analysis module, which is used to select an intelligent algorithm model that matches the application scenario and transaction object of the identified transaction data from an intelligent algorithm library according to the application scenario and transaction object of the identified transaction data, and analyze the preprocessed transaction data using the matched intelligent algorithm model to obtain a transaction data analysis result; the intelligent algorithm library stores a number of intelligent algorithms generated by training using training data containing historical transaction data analysis results of different application scenarios and transaction objects, as well as a mapping table of historical application scenarios and transaction objects and intelligent algorithms; The transaction data storage and display module is used to store the transaction data analysis results obtained based on different transaction objects and display them visually to the corresponding transaction terminals. By adopting the above solution, we can achieve real-time collection of transaction data, efficient preprocessing, adaptive intelligent analysis and visual display of corresponding items of user needs. In a third aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the method as described above.

[0021] In a fourth aspect, the present application provides a computer device, comprising a memory, a processor, and a program stored and executable on the memory, wherein the program implements the steps of the above method when executed by the processor.

[0022] In summary, this application has the following beneficial effects: 1. By using edge computing devices with built-in lightweight AI models to pre-process the collected transaction data, the real-time and accuracy of data processing can be significantly improved, effectively avoiding the delay problem that is prone to occur in the data pre-processing link of traditional methods; using machine learning technology to automatically identify the application scenarios and transaction objects of pre-processed transaction data, and selecting intelligent algorithm models that match the current transaction data from the intelligent algorithm library for analysis, the flexibility and adaptability of data processing are enhanced; the transaction data analysis results obtained based on different transaction objects are stored and visualized on the corresponding transaction terminals, which improves the transparency of the transaction market and user experience; 2. Considering the comprehensiveness of transaction data analysis, the knowledge topology of the transaction platform is used to effectively collect and standardize the cross-platform transaction data of strongly related transaction platforms, which improves the accuracy and efficiency of transaction data collection. 3. Select the appropriate AI model according to the application scenario of the transaction data to further ensure the efficiency of the preprocessing process, and automatically switch to other AI models when the computing resources of the determined edge computing device are insufficient, thereby ensuring the stability of the preprocessing process and the integrity of the transaction data analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a flow chart of the data processing method of the intelligent algorithm described in the specific embodiment; Figure 2 It is a structural diagram of the data processing system of the intelligent algorithm described in the specific embodiment. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0025] like Figure 1 As shown, the embodiment of the present application discloses a data processing method of an intelligent algorithm, and the specific steps include: S1. Collect transaction data in real time.

[0026] Specifically, the transaction data may be collected by using, but not limited to, API interfaces opened by various trading platforms or other collection methods, such as obtaining data through a real-time data transmission interface implemented by Web Socket, or obtaining data through a data push mode implemented by a subscription and publishing mechanism.

[0027] Taking into account the comprehensiveness and accuracy of transaction data analysis, the latest knowledge topology map of each trading platform is requested from each trading platform through the HTTP or HTTPS protocol, and the graph analysis method is used to obtain the correlation score of each trading platform in the knowledge topology map, and the trading platforms with correlation scores greater than the preset scores are obtained, that is, the strongly correlated trading platforms in the knowledge topology map, and the strongly correlated trading platforms in the knowledge topology map are taken as the trading platforms to be collected.

[0028] The preset score is a fixed threshold preset by humans, such as 0.8, or is set as a dynamic threshold, which is automatically adjusted according to the average score of the correlation of historical transaction data. The correlation score of each trading platform in the knowledge topology graph is obtained by using graph analysis, which can be calculated by matrix operations, such as using the adjacency matrix of the graph for calculation or by path length analysis to evaluate the correlation strength between different trading platforms, such as Dijkstra algorithm calculation.

[0029] The API interfaces of each trading platform are used to complete the acquisition of cross-platform transaction data of strongly related trading platforms, and the collected transaction data is converted and exchanged according to the standardized data format in the cross-platform transaction data sharing agreement. Among them, the format conversion can use a common data format conversion tool, such as Apache Avro, and the data exchange follows the sharing agreement to ensure the consistency and security of data between different platforms.

[0030] S2. Use machine learning technology to automatically identify the application scenarios and transaction objects of pre-processed transaction data.

[0031] Specifically, a machine learning model for identifying application scenarios and transaction objects of transaction data is constructed, which is recorded as a scenario object recognition model; the input of the model is transaction data, and the output of the model is the application scenarios and transaction objects of transaction data, and it is trained by historical data annotated with transaction scenarios and transaction objects. This embodiment is explained based on ticket transaction data. The application scenarios of the ticket transaction data include normal transactions, peak-time transactions, and promotional transactions; the transaction objects include ticket buyers, ticket sellers, and third-party resellers. Using the constructed machine learning model, the application scenarios and transaction objects of the currently collected transaction data can be identified.

[0032] S3. Use edge computing devices with built-in lightweight AI models to pre-process transaction data.

[0033] Specifically, due to the large amount of transaction data, edge computing devices are chosen to preprocess the transaction data. The preprocessing includes: abnormal data cleaning, missing data compensation, etc. In order to further achieve efficient and accurate preprocessing of transaction scenarios, edge computing devices with built-in lightweight AI models are selected to complete the preprocessing of transaction data.

[0034] Among them, the built-in lightweight AI models can be selected and set according to different application scenarios, such as neural networks, decision trees, etc. Some built-in lightweight AI models are generated by training the transaction data in a fixed application scenario after preprocessing; for example, for edge computing devices with rich computing resources, a lightweight AI model in the promotion scenario is selected to preprocess the transaction data in the promotion scenario.

[0035] In addition, in order to ensure the efficiency of transaction data preprocessing, it is also necessary to ensure the stability of transaction data preprocessing. Therefore, according to the application scenarios of the identified transaction data, the edge computing devices with the same application scenarios as the transaction data are determined among the edge computing devices. The number of collected transaction data is counted to determine whether the computing resources of the determined edge computing device can meet the preprocessing of the current transaction data volume. If not, the remaining AI models are used for preprocessing.

[0036] S4. Select a matching intelligent algorithm model from the intelligent algorithm library, use the matching intelligent algorithm model to analyze the preprocessed transaction data, and obtain the transaction data analysis results.

[0037] Specifically, an intelligent algorithm library is constructed, which stores a number of intelligent algorithms generated by training with training data containing historical transaction data analysis results of different application scenarios and transaction objects, and a mapping table of historical application scenarios and transaction objects and intelligent algorithms. The transaction data analysis results containing historical different application scenarios and transaction objects can be obtained through expert analysis and based on user demand analysis.

[0038] Considering that different transaction objects have different needs for transaction data analysis, for example, the ticket seller pays more attention to analyzing user preferences through transaction data to make accurate customer recommendations, while the ticket buyer pays more attention to analyzing the probability of successful ticket purchases at the moment and recommending ticket purchase times and plans to avoid peak ticket purchases through transaction data; both the ticket seller and the ticket buyer pay attention to abnormal transaction monitoring and analysis of transaction data to ensure transaction security. It can be seen that the intelligent algorithm models in the self-intelligent algorithm library include: intelligent algorithm models that match the application scenarios of the identified transaction data and unilateral transaction objects, and intelligent algorithm models that match the application scenarios of the identified transaction data and multi-party transaction objects.

[0039] According to the application scenario and transaction object of the identified transaction data, an intelligent algorithm model that matches the application scenario and transaction object of the identified transaction data is selected from the intelligent algorithm library; for example: if the currently identified transaction data are all ticket purchase application data sent by the ticket purchaser, and the current application scenario is peak-period transaction, the first intelligent algorithm model that matches the peak-period transaction application scenario and the transaction object of the ticket purchaser is obtained; if the currently identified transaction data are ticket purchase transaction data completed between the ticket purchaser and the transaction party, and the current application scenario is a normal transaction, the second intelligent algorithm model that matches the normal transaction application scenario and multi-party transaction objects can be obtained, and the third intelligent algorithm model that matches the normal transaction application scenario and the transaction object of the ticket purchaser and the fourth intelligent algorithm model that matches the normal transaction application scenario and the transaction object of the ticket seller can be obtained.

[0040] The pre-processed transaction data is analyzed using a matched intelligent algorithm model to obtain a transaction data analysis result; wherein, the analysis of the pre-processed transaction data by each matched intelligent algorithm model is independent and does not interfere with each other.

[0041] S5. The transaction data analysis results obtained based on different transaction objects are stored separately and displayed visually to the corresponding transaction terminals.

[0042] Specifically, the transaction data analysis results obtained based on different transaction objects are stored separately; for example, the transaction data analysis results for the purchaser, the ticket seller, and the third-party reseller are stored separately, and the corresponding transaction data analysis results are displayed as needed.

[0043] In a specific embodiment, in order to further improve the accuracy of data analysis, the model parameters can be adjusted dynamically based on the characteristic changes of the transaction data of the intelligent algorithm set. The method also includes: The features of the preprocessed transaction data are extracted using a matching intelligent algorithm model; wherein the features of the transaction data include data volume, data distribution, and transaction speed, etc.

[0044] According to the extracted features of the preprocessed transaction data, the corresponding intelligent algorithm model parameters are matched to complete the parameter adjustment of the intelligent algorithm model, and the analysis of the preprocessed transaction data is continued with the intelligent algorithm model after parameter adjustment.

[0045] Among them, the intelligent algorithm model parameters for feature matching of the extracted preprocessed transaction data are the intelligent model parameters with the highest user satisfaction score in the analysis results obtained by the intelligent algorithm model corresponding to the intelligent algorithm model parameters of historical matching, that is, the historical transaction data are analyzed using intelligent algorithm models corresponding to different intelligent algorithm model parameters to obtain user satisfaction scores for the analysis results, and the intelligent algorithm parameters with the highest satisfaction scores are obtained as the intelligent algorithm parameters for feature matching of the corresponding historical transaction data.

[0046] In a specific embodiment, considering that each transaction party has different requirements for the time required to obtain the transaction data analysis results, the matched intelligent algorithm model is correspondingly screened to quickly display the transaction data analysis results. The method further includes: Receive the time requirement for obtaining data analysis results sent by the transaction object; According to the average time it takes for each matching intelligent algorithm model to analyze the transaction data under the same historical application scenario and transaction object scenario to obtain the transaction data analysis results, the intelligent algorithm model whose average time for obtaining the transaction data analysis results is longer than the required acquisition time is selected from the currently matched intelligent algorithm models, and multiple matching results are calculated using the retained matching multiple intelligent algorithm models.

[0047] In addition, in order to meet the transaction object's demand for obtaining typical data analysis results, the method also includes: Receive requests from transaction partners for obtaining typical data analysis results; It is determined whether there is an intelligent algorithm model whose output is a typical transaction data analysis result from the currently matched intelligent algorithm model. If not, a prompt message indicating that the typical transaction data analysis result cannot be obtained is output and fed back to the trading terminal for corresponding visual display. Otherwise, the intelligent algorithm model whose output is a typical transaction data analysis result is retained from the currently matched intelligent algorithm model, and the transaction data is analyzed using the retained intelligent algorithm model whose output is a typical transaction data analysis result. The typical transaction data analysis result includes: abnormal transaction data analysis result.

[0048] In a specific embodiment, considering that there may be analysis timeout or failure when using intelligent algorithm models for data analysis, in order to accurately determine the cause and take appropriate measures to continue to complete the data analysis, the efficiency and accuracy of the data analysis are guaranteed; the method also includes: When a timeout or failure occurs when analyzing the preprocessed transaction data using an intelligent algorithm model that matches the application scenario of the identified transaction data and multi-party transaction objects, check whether an intelligent algorithm model that matches the application scenario of the identified transaction data and a unilateral transaction object is also matched and whether a timeout or failure occurs when analyzing the data using the intelligent algorithm model that matches the application scenario of the identified transaction data and the unilateral transaction object. If a timeout or failure occurs again, it is determined that there is a problem with the corresponding unilateral transaction object data, and the transaction data is re-preprocessed, and then data analysis is performed on the re-preprocessed transaction data; otherwise, a spare intelligent algorithm model that matches the application scenario of the identified transaction data and multi-party transaction objects is directly selected to complete the analysis of the transaction data.

[0049] Similarly, when a timeout or failure occurs when analyzing the preprocessed transaction data using an intelligent algorithm model that matches the application scenario of the identified transaction data and a single-party transaction object, check whether an intelligent algorithm model that matches the application scenario of the identified transaction data and multiple transaction objects is also matched and whether a timeout or failure occurs when analyzing the data using the intelligent algorithm model that matches the application scenario of the identified transaction data and multiple transaction objects. If a timeout or failure occurs again, it is determined that there is a problem with the corresponding multiple transaction object data, and the transaction data is re-preprocessed, and then data analysis is performed on the re-preprocessed transaction data; otherwise, directly select a backup intelligent algorithm model that matches the application scenario of the identified transaction data and multiple transaction objects to complete the analysis of the transaction data.

[0050] In a specific embodiment, considering that different transaction objects have the need to obtain transaction data analysis results of the transaction objects for designated transaction objects other than themselves, in order to controllably share data between transaction objects, data security and privacy protection are improved; the method further includes: Obtain a request from a transaction object for obtaining transaction data analysis results of a designated transaction object other than itself, and determine whether the current transaction object has a preset permission to obtain the transaction data analysis results of the designated transaction object; wherein the preset permission is a permission pre-set by a transaction platform or user terminal corresponding to each transaction object to allow a transaction platform or user terminal corresponding to a specific transaction object to view its data analysis results.

[0051] If it exists, the transaction data analysis result of the specified transaction object will be returned to the transaction object terminal that sent the acquisition request.

[0052] Obtain the transaction object's request for sharing its own transaction data analysis results, determine whether the current transaction object has the preset activation authority to share its own transaction data analysis results, and if so, transmit the transaction object's own transaction data analysis results to all transaction object terminals for display. In addition, after determining that the current transaction object has the preset activation authority to share its own transaction data analysis results, a prompt message for confirming sharing can be sent, and after obtaining the confirmation message of the transaction object confirming sharing, the transaction object's own transaction data analysis results can be transmitted to all transaction object terminals for display, thereby further ensuring the security of the shared transaction data analysis results.

[0053] like Figure 2 As shown, the embodiment of the present application discloses a data processing system based on an intelligent algorithm, comprising: The transaction data collection module 101 is used to collect transaction data in real time; A transaction data classification module 102, for automatically identifying application scenarios and transaction objects of pre-processed transaction data using machine learning technology; The transaction data processing module 103 is used to pre-process the collected transaction data using an edge computing device with a built-in lightweight AI model to obtain the pre-processed transaction data; The transaction data analysis module 104 is used to select an intelligent algorithm model that matches the application scenario and transaction object of the identified transaction data from the intelligent algorithm library according to the application scenario and transaction object of the identified transaction data, and analyze the preprocessed transaction data using the matched intelligent algorithm model to obtain transaction data analysis results; the intelligent algorithm library stores a number of intelligent algorithms generated by training using training data containing historical transaction data analysis results of different application scenarios and transaction objects, as well as a mapping table between historical application scenarios and transaction objects and intelligent algorithms; The transaction data storage and display module 105 is used to store the transaction data analysis results obtained based on different transaction objects and to display them visually to the corresponding transaction terminals.

[0054] In a specific embodiment, the transaction data analysis module 104 in the system is also used to check whether an intelligent algorithm model matching the application scenario of the identified transaction data and the unilateral transaction object is matched at the same time when analyzing the pre-processed transaction data. And whether a timeout or failure occurs when using the intelligent algorithm model matching the application scenario of the identified transaction data and the unilateral transaction object for analysis. If a timeout or failure also occurs, it is determined that there is a problem with the corresponding unilateral transaction object data, and the transaction data is re-preprocessed. Otherwise, a spare intelligent algorithm model matching the application scenario of the identified transaction data and the multi-party transaction objects is directly selected to complete the analysis of the transaction data.

[0055] In a specific embodiment, the transaction data storage and display module 105 is also used to obtain a request from a transaction object for obtaining transaction data analysis results of a specified transaction object other than itself, and to determine whether the current transaction object has a preset activation authority to obtain the transaction data analysis results of the specified transaction object; if so, the transaction data analysis results of the specified transaction object are transmitted back to the transaction object terminal that sent the acquisition request.

[0056] In a specific embodiment, the transaction data storage and display module 105 is also used to obtain a request from a transaction object to share its own transaction data analysis results, determine whether the current transaction object has a preset activation authority to share its own transaction data analysis results, and if so, transmit the transaction data analysis results of the transaction object itself to all transaction object terminals for display.

[0057] The embodiment of the present application also discloses a computer-readable storage medium.

[0058] Specifically, the computer-readable storage medium stores a computer program that can be loaded by a processor and execute a data processing method such as the above-mentioned intelligent algorithm. The computer-readable storage medium includes, for example: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0059] The embodiment of the present application also discloses a computer device.

[0060] Specifically, the computer device includes a memory and a processor, and the memory stores a computer program that can be loaded by the processor and execute the data processing method of the above-mentioned intelligent algorithm.

[0061] The above are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Any feature disclosed in this specification (including the abstract and drawings), unless otherwise stated, can be replaced by other equivalent or alternative features with similar purposes. That is, unless otherwise stated, each feature is only an example of a series of equivalent or similar features.

Claims

1. A data processing method based on intelligent algorithm, characterized in that: include: Collect transaction data in real time; Use machine learning technology to automatically identify the application scenarios and transaction objects of pre-processed transaction data; Use edge computing devices with built-in lightweight AI models to pre-process the collected transaction data and obtain pre-processed transaction data; According to the application scenarios and transaction objects of the identified transaction data, an intelligent algorithm model matching the application scenarios and transaction objects of the identified transaction data is selected from the intelligent algorithm library, and the pre-processed transaction data is analyzed using the matching intelligent algorithm model to obtain transaction data analysis results; the intelligent algorithm library stores a number of intelligent algorithms generated by training using training data containing historical transaction data analysis results of different application scenarios and transaction objects, as well as a mapping table of historical application scenarios and transaction objects and intelligent algorithms; The transaction data analysis results obtained based on different transaction objects are stored separately and displayed visually to the corresponding transaction terminals.

2. The data processing method based on intelligent algorithm according to claim, characterized in that: The real-time collection of transaction data specifically includes: The knowledge topology map of each trading platform is obtained, and the correlation score of each trading platform in the knowledge topology map is obtained by using the graph analysis method, and then each trading platform with a correlation score greater than the preset score is obtained as a strongly correlated trading platform in the knowledge topology map; the API interface of each trading platform is used to complete the acquisition of cross-platform transaction data of strongly correlated trading platforms, and the collected transaction data is converted and exchanged in accordance with the standardized data format in the cross-platform transaction data sharing agreement.

3. The data processing method based on intelligent algorithm according to claim, characterized in that: The edge computing device with a built-in lightweight AI model performs abnormal data cleaning and preprocessing on the collected transaction data, specifically including: The edge computing device with a built-in lightweight AI model may adopt different AI models, including some AI models generated by training transaction data in a fixed application scenario after preprocessing; determining the edge computing device with the same application scenario as the transaction data in each edge device according to the application scenario of the identified transaction data; counting the number of collected transaction data, and determining whether the computing resources of the determined edge computing device meet the preprocessing of the current transaction data volume. If not, using the remaining AI models for preprocessing.

4. The data processing method based on intelligent algorithm according to claim 1, characterized in that: The use of the matching intelligent algorithm model to analyze the pre-processed transaction data also includes: Use the matching intelligent algorithm model to extract the characteristics of pre-processed transaction data, including data volume, data distribution and speed; According to the intelligent algorithm model parameters corresponding to the feature matching of the extracted preprocessed transaction data, the parameter adjustment of the intelligent algorithm model is completed, and the analysis of the preprocessed transaction data is continued with the intelligent algorithm model after parameter adjustment; the intelligent algorithm model parameters for feature matching of the extracted preprocessed transaction data are the intelligent model parameters with the highest user satisfaction score in the analysis results obtained by the intelligent algorithm model corresponding to the intelligent algorithm model parameters matched according to the historical matching.

5. The data processing method based on intelligent algorithm according to claim, characterized in that: Also includes: The intelligent algorithm model selected from the intelligent algorithm library to match the application scenario and transaction object of the identified transaction data includes an intelligent algorithm model matching the application scenario of the identified transaction data and a unilateral transaction object, and an intelligent algorithm model matching the application scenario of the identified transaction data and a multi-party transaction object; When a timeout or failure occurs when analyzing the pre-processed transaction data using an intelligent algorithm model that matches the application scenario of the identified transaction data and multi-party transaction objects, check whether an intelligent algorithm model that matches the application scenario of the identified transaction data and a unilateral transaction object is also matched and whether a timeout or failure occurs when analyzing the data using the intelligent algorithm model that matches the application scenario of the identified transaction data and the unilateral transaction object. If a timeout or failure occurs again, it is determined that there is a problem with the corresponding unilateral transaction object data, and the transaction data is preprocessed again. Otherwise, a spare intelligent algorithm model that matches the application scenario of the identified transaction data and multi-party transaction objects is directly selected to complete the analysis of the transaction data.

6. The data processing method based on intelligent algorithm according to claim, characterized in that: Also includes: Obtain the transaction data analysis results of the specified transaction objects other than itself by the transaction object, and determine whether the current transaction object has the preset activation authority to obtain the transaction data analysis results of the specified transaction object. If so, the transaction data analysis results of the specified transaction object are returned to the transaction object terminal that sent the acquisition request.

7. The data processing method based on intelligent algorithm according to claim, characterized in that: Also includes: Obtain the transaction object's request to share its own transaction data analysis results, determine whether the current transaction object has the preset activation authority to share its own transaction data analysis results, and if so, transmit the transaction object's own transaction data analysis results to all transaction object terminals for display.

8. A data processing system based on intelligent algorithm, characterized in that: include: Transaction data collection module, used to collect transaction data in real time; The transaction data classification module is used to automatically identify the application scenarios and transaction objects of the pre-processed transaction data using machine learning technology; A transaction data processing module is used to pre-process the collected transaction data using an edge computing device with a built-in lightweight AI model to obtain the pre-processed transaction data; A transaction data analysis module, which is used to select an intelligent algorithm model that matches the application scenario and transaction object of the identified transaction data from an intelligent algorithm library according to the application scenario and transaction object of the identified transaction data, and analyze the preprocessed transaction data using the matched intelligent algorithm model to obtain a transaction data analysis result; the intelligent algorithm library stores a number of intelligent algorithms generated by training using training data containing historical transaction data analysis results of different application scenarios and transaction objects, as well as a mapping table of historical application scenarios and transaction objects and intelligent algorithms; The transaction data storage and display module is used to store the transaction data analysis results obtained based on different transaction objects and to display them visually to the corresponding transaction terminals.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 7.

10. A computer device, characterized in that: The computer device comprises a memory, a processor and a program stored and executable on the memory, and the program implements the steps of the method according to any one of claims 1 to 7 when executed by the processor.

Images