Detailed Description
According to the data flow supervision and early warning method for the financing service, the problems that in the prior art, integration of data of different banks and company main bodies is difficult, comprehensive supervision of comprehensive financing service cannot be achieved, delay exists in the aspect of processing real-time flow data, analysis and monitoring cannot be conducted timely, and timeliness is poor, efficiency is low and accuracy is low are solved.
Having described the basic principles of the present application, various non-limiting embodiments of the present application will now be described in detail with reference to the accompanying drawings.
Example 1
As shown in fig. 1, an embodiment of the present application provides a data flow supervision and early warning method for financing service, where the method includes:
establishing a general data set, wherein the general data set is obtained through an interactive bank database, and after communication connection with the bank database is established, reading company main body data and bank main body data construction of financing service;
establishing communication connection with a bank database, and after the connection is successful, executing query operation to read company main body data related to financing business, including company name, registration information, legal person and the like; another query operation is performed to obtain bank principal data including bank names, license information, etc. And integrating the acquired company main body data with the bank main body data to acquire a general data set, and performing data cleaning and preprocessing on the data in the general data set, wherein the operations comprise removing repeated items, processing missing data, standardizing data formats and the like so as to ensure the consistency and quality of the data. And storing the processed data set to accurately reflect financing service and bank information.
Configuring a node update database of the edge computing node, and generating an edge computing node matching network by taking the node update database as matching data;
edge computing aims to handle computing tasks close to the data sources to reduce the burden on the central data center and reduce data transmission delays. The edge computing node is configured with a node update database for storing relevant data such as node information, states and the like, and the database can record the information such as node update, state change and the like so as to facilitate monitoring and management. The information in the node update database is used as matching data, which means that when it is desired to select the best edge computing node for a particular computing task, these matching data can be referenced, e.g., the node that is most suitable for performing the particular task is selected based on factors such as the current load of the node, geographical location, available resources, etc. And using the nodes to update the matching data in the database to generate an edge computing node matching network, wherein the network determines which edge computing node is allocated with the computing task through the weight, the distance and the like of the nodes. This ensures that the computing tasks can be handled in the best location and environment, thereby optimizing overall computing performance and efficiency.
Obtaining a data flow supervision record of a current financing service, extracting the data flow supervision record, and constructing a node selection bias factor based on an extraction result;
communication is established with the relevant regulatory authorities, financial institutions or data providers, and data flow monitoring records related to the current financing service are acquired, wherein the records comprise information such as histories of financing transactions, time stamps of data transmission, transaction amounts, quality indexes and the like. From the obtained regulatory records, operations such as data parsing, screening, conversion, aggregation and the like are performed to identify and extract key information.
The process of constructing the node selection bias factor comprises the steps of carrying out trend analysis and key feature extraction on the extracted data, wherein the trend analysis comprises the steps of detecting the growth trend, the periodic fluctuation, the abnormal value and the like of the data, so as to know the dynamic change of the data; key feature extraction aims at identifying key features related to financing services, including identifying common transaction features, data quality features, key service indicators, etc., to determine which factors are most important for node selection.
Based on the results of trend analysis and key feature extraction, a node selection bias factor, which may be a weight, score, or other evaluation criteria, is constructed to measure the importance and applicability of different edge computing node attributes, such as, for example, computing power, network delay, reliability, etc., of the node, and illustratively, if the data traffic trend is significant, a node that is capable of handling a large amount of data is more prone to be selected.
According to the actual situation, the constructed node selection bias factors are subjected to weight adjustment and verification, for example, the weight is evaluated by experts and technical teams in related fields, and the adjusted weight can accurately reflect the preference of node selection and is matched with the actual data flow supervision record.
By constructing the node selection bias factors, the node selection process can be guided more accurately, which is an automatic decision process, so that tasks can be executed on the most suitable nodes, and further, the resource allocation and task scheduling of the edge computing network are optimized, and the system performance and efficiency are improved.
Reading real-time circulation data, generating a data feature set of the real-time circulation data, and transmitting the data feature set, the node selection bias factor and the general data set to the edge computing node matching network by taking the data feature set, the node selection bias factor and the general data set as configuration data to finish the selection and initialization of computing nodes;
by establishing connection with corresponding data sources, such as a bank database, real-time circulation data of the current financing service are inquired and acquired, the data come from real-time information of transaction or other service processes, and for the acquired real-time circulation data, a statistical method is used for carrying out feature extraction operation, such as calculating features of transaction amount, transaction frequency, participant information and the like, according to specific service requirements and supervision requirements, and the features are used as part of a data feature set. After extracting features, the extracted features are combined into a data feature set, which is a structured data set, wherein each sample represents a real-time stream of data and each feature corresponds to a certain attribute or index of the sample.
Preparing an edge computing node matching network, wherein the network comprises a position matching sub-network, a service adaptation sub-network and an integration sub-network, transmitting configuration data consisting of a data feature set, a node selection deflection factor and a general data set to the edge computing node matching network through network connection, adjusting and compensating network parameters in the edge computing node matching network through network parameter compensation, for example, weighing applicability of different nodes based on the node selection deflection factor, and compensating the network parameters of the edge computing node matching network so as to ensure that tasks are distributed to the most suitable nodes for execution, thereby improving the selection accuracy and suitability of the nodes.
Executing data analysis corresponding to the real-time circulation data through the initialized computing node to generate a first analysis result;
the edge computing node matching network comprises a position matching sub-network, a service adaptation sub-network and an integration sub-network, wherein the position matching sub-network is used for matching proper edge computing nodes according to position information in configuration data, the sub-network can use a Geographic Information System (GIS) technology or other position identification methods to compare and match position characteristics in the configuration data with positions of the edge computing nodes, and edge computing nodes which are closer to a data source can be found through position matching so as to reduce data transmission delay and network congestion;
The service adaptation sub-network is used for carrying out service adaptation matching on the edge computing nodes according to the service requirements in the configuration data and the general data set, the sub-network is extracted according to the service characteristics in the configuration data and carries out matching calculation with the weights in the general data set so as to determine which edge computing nodes are most suitable for processing specific service requirements, and the utilization of computing resources and the efficiency of service processing can be optimized through service adaptation;
the integration sub-network is responsible for normalizing and integrating the results of the position matching sub-network and the service adaptation sub-network, and the sub-network relates to normalization calculation of the results so as to ensure that the matching results generated by different sub-networks have consistent measurement standards, and the final node selection can be obtained through the processing of the integration sub-network.
The sub-networks work together, select proper edge computing nodes through position matching, service adaptation and integration processing, provide optimal computing resources for subsequent tasks or data processing, and aim at finding the edge computing nodes capable of meeting data feature sets and node selection bias factors according to requirements in configuration data, and realize optimal allocation of the computing resources.
Further, the edge computing node matching network includes a location matching sub-network, a traffic adaptation sub-network, and an integration sub-network, the method further comprising:
extracting data position features in the data feature set, and sending the data position features to the position matching sub-network;
performing position matching calculation of edge calculation nodes on the data position features through the position matching sub-network to generate a position matching result;
extracting service characteristics of the data characteristic set through the service adaptation sub-network, and carrying out weight service adaptation matching on the extraction result and the general data set to generate a service matching result;
and sending the position matching result and the service matching result to the integration sub-network, executing normalization processing, and obtaining the first analysis result according to the processing result.
The structure and fields of the data feature set are examined from the data feature set to find a corresponding feature column containing location information, which is extracted from the selected location feature column by parsing the address text, extracting latitude and longitude coordinates, or using other geocoding techniques. When the location information is successfully extracted, it is converted into an appropriate format for transmission to the location matching sub-network, for example, if the location information is represented in latitude and longitude coordinates, it may be converted into a character string form having a uniform standard. And sending the extracted and converted data position characteristics to a position matching sub-network for processing.
In the location matching sub-network, the data location characteristics are compared with the location information of the edge computing nodes, distance measures between the data location characteristics and the edge computing nodes are calculated, the distance measures are ordered from near to far according to the result of the location matching calculation, and a location matching result is generated, which can be a ranking list, the location matching degree of each edge computing node is displayed from high to low, or an identifier of the best matching node is directly indicated, so that transmission delay and network congestion are reduced as much as possible in the data processing process.
Features related to the business are extracted from the data feature set by adopting feature extraction technologies such as text analysis, image processing and feature engineering. According to the extracted service characteristics and the general data set, weight service adaptation matching is executed in the service adaptation sub-network, specifically, the service characteristics are compared with key characteristics in the general data set, the similarity between the service characteristics and the key characteristics is calculated, the similarity calculation result is used as a result of the weight service adaptation matching, a service matching result is generated, and likewise, the service matching result can be a ranking list, the service matching degree of each edge calculation node is displayed, or the identifier of the best matching node is directly indicated.
And receiving the position matching result and the service matching result from the position matching sub-network and the service matching sub-network, normalizing the position matching result and the service matching result, which means that the results of different ranges, units or measurement modes are unified into the same standard range or unit, and performing combination processing, such as weighted summation, multiplication operation and the like, according to the normalized position matching result and the normalized service matching result to obtain an integrated processing result, and obtaining a first analysis result based on the combined processing result.
Further, the method further comprises:
performing network parameter compensation on the edge computing node matching network based on the node selection bias factor;
after the position matching result and the service matching result are sent to the integration sub-network, performing adaptive reconstruction of the position matching result and the service matching result through a network parameter compensation result;
and (5) finishing normalization processing by adapting the reconstruction result.
The compensation of network parameters based on node selection bias factors is an optimization strategy to better accommodate the needs of node selection. Specifically, according to the node selection bias factor, the value of the network parameter is adjusted through weighted calculation so as to better reflect the preference of node selection, for example, in the case that the node selection bias calculation capability is higher, the weight of the node with higher calculation capability can be increased or the related parameter can be adjusted so as to improve the importance of the node in the network. In this way, the need for node selection can be better met, enabling the network to more accurately match and select appropriate edge computing nodes, which helps to improve overall performance and efficiency, and optimize utilization of system resources.
According to the network parameter compensation process, network parameter compensation is performed in the integrated sub-network, including adjusting the weights of the network parameters to better accommodate the requirements of location matching and service matching. Using the network parameter compensation result to adapt the position matching result to reconstruct, i.e. recalculate the position matching score or adjust the weight of the position matching result, so as to more accurately reflect the node selection preference; likewise, using the network parameter compensation results, adapting the traffic matching results to reconstruct, including recalculating the traffic matching scores or adjusting the weights of the traffic matching results, to better meet the traffic demands. And outputting the position matching result and the service matching result after the reconstruction, which are suitable for the reconstruction, are favorable for improving the selection accuracy and improve the overall system performance according to the optimization result of the network parameters.
The position matching result and the service matching result after the reconstruction are obtained from the integrated sub-network, and the required normalization range is determined according to specific requirements, which can be a predefined standard range or determined according to data analysis and service requirements. Factors for normalization processing are calculated from the appropriate reconstruction results and target ranges, for example, using common normalization methods such as min-max normalization, Z-score normalization, and the like.
And applying the calculated normalization factor to an adaptive reconstruction result, normalizing the position matching result and the service matching result to obtain the normalized position matching result and service matching result, wherein the normalized position matching result and service matching result have the same standard range or unit, and are convenient for subsequent comparison and analysis. This helps eliminate dimensional differences between the results and provides more consistent and comparable data for further analysis and decision making.
Extracting key monitoring data of the computing nodes, sending the key monitoring data to a processing center, and analyzing the key monitoring data of all the computing nodes through the processing center to generate a second analysis result;
and extracting key monitoring data from each computing node, and sending the key monitoring data to a processing center through network communication. A data analysis method is selected, for example, using cluster analysis to establish a group pattern between computing nodes, using anomaly detection methods to identify anomalous nodes, or using time series analysis to predict node performance. Under the selected data analysis method, key monitoring data analysis is carried out, wherein the key monitoring data analysis comprises the steps of calculating the average value, variance and the like of the nodes by using statistics, constructing a prediction model by using a machine learning algorithm, or finding hidden modes and rules by using a data mining technology. The results of the analysis are presented by visual methods such as graphs, thermodynamic diagrams, etc. to support understanding of the overall performance, trends, and anomalies of the computing node population. Based on the data analysis and display process, a second analysis result is generated, so that the performance, stability and safety of the computing node can be deeply known, and decision support is provided for data flow supervision and early warning of financing service.
Further, the method further comprises:
establishing an interest feature set of each data flow node, wherein the interest feature set is obtained by analyzing financing service;
when data analysis of corresponding real-time circulation data is carried out through a computing node, data extraction in a preset period is carried out through the interest feature set, and the data is recorded as a first data set;
if the data extraction result in any preset period is a null result, reselecting the mean characteristic data in the preset period, and marking the selection result as a second data set;
the critical monitoring data is generated from the first data set and the second data set.
Data flow nodes refer to specific links involved in data transfer and processing in the whole financing process, the nodes comprise banks, enterprises, partners, regulatory authorities and the like, and interest characteristics of the data flow nodes are defined according to specific financing business requirements and targets, for example, for edge computing nodes, the interest characteristics comprise processing capacity, network bandwidth, energy consumption and the like of the nodes.
In order to establish the interest feature set, the financing service needs to be parsed, specifically, various data related to the financing service, including transaction records, transaction documents, contract information, etc., are collected, data analysis is performed to determine which data attributes are critical to data processing, and based on the results of the data analysis, an interest feature set is established for each data flow node, including the specific data attributes of interest to that node, which feature sets differ from node to node because different nodes focus on different data aspects.
In financing services, real-time streaming data is continuously generated data, such as transaction records, contract information, etc., which are responsible for analysis by the computing nodes to obtain information about the status of the service. The predetermined period is set according to actual conditions and specific requirements, and means that data extraction operation is performed in real-time circulation data according to a time range of the predetermined period at certain time intervals, such as every hour and every day, by using an interest feature set, and according to the interest features defined in the interest feature set, data items meeting the conditions are screened out, and the extracted data items are combined into a first data set, wherein the first data set contains information related to the interest features in the predetermined period and is used for subsequent analysis and monitoring.
After the data extraction in the preset period is completed, the extraction result is checked, if the extraction result is empty, the real-time circulation data which accords with the interesting characteristic condition does not exist in the period, and the data needs to be reselected. Specifically, in the predetermined period, the average value of each feature in the interest feature set is obtained by performing aggregate calculation on all data in the period, based on the calculated average value feature data, the data items related to the interest feature are reselected, the data items may represent the average condition in the whole period, the reselected average value feature data are combined into a second data set, and the data set may be compared with the first data set or used alone to meet further analysis and requirements. This ensures that even if there is no eligible real-time streaming data for a certain period, there is still data available for subsequent analysis or decision making.
And merging the first data set and the second data set as the key monitoring data so as to fully utilize the information of the real-time data and the historical data.
And carrying out data flow supervision and early warning on financing service through the first analysis result and the second analysis result.
Further, the method comprises the steps of:
carrying out overall anomaly analysis through the second analysis result, determining overall anomaly trend, and generating auxiliary recognition factors for local monitoring;
analyzing the first analysis result to obtain a local abnormal evaluation result, and optimizing a matching result of the local abnormal evaluation result through the auxiliary recognition factor;
and finishing the data flow supervision and early warning of the financing service through the optimized result.
And carrying out overall anomaly analysis on the whole system by utilizing the collected second analysis result, wherein the aim is to determine overall anomaly trend, namely finding anomaly condition and trend in the whole system, for example, calculating average value, variance and the like of nodes according to the second analysis result, predicting data trend and the like, carrying out anomaly node identification, and generating auxiliary identification factors for local monitoring according to the overall anomaly analysis result, wherein the auxiliary identification factors can be a series of indexes for assisting subsequent local anomaly evaluation, helping to capture potential anomaly condition or providing more accurate anomaly identification capability.
Analyzing the result of the first analysis, extracting information related to local abnormal evaluation, wherein the information can be abnormal condition description about calculation nodes, indexes, parameters or other related data, and calculating, comparing or evaluating certain indexes or rules to determine whether local abnormal conditions exist or not so as to obtain a local abnormal evaluation result.
The generated auxiliary recognition factors are applied to the local abnormal evaluation results and matched with the local abnormal evaluation results to further judge and recognize abnormal conditions and provide more accurate evaluation results, and the local abnormal evaluation results are optimized according to the matching results of the auxiliary recognition factors, including further filtering, correction or adjustment, so that the evaluation results are more accurate, reliable and suitable for actual needs.
Through the steps, the identification and evaluation capability of local abnormal conditions is improved, and the abnormal conditions in the financing business data circulation process can be monitored more carefully.
Based on the analysis result and the auxiliary recognition factor, the data flow of the financing service is optimized, supervision and early warning are carried out, which comprises setting a reasonable threshold value so as to timely detect potential abnormal behaviors or risks in the monitoring process, triggering corresponding early warning prompts, and timely conveying early warning information to related personnel so that the related personnel can take necessary measures to treat and manage potential problems, thus realizing effective supervision and early warning on the data flow of the financing service.
Further, the method further comprises:
judging whether the auxiliary recognition factor meets a preset threshold value or not;
if the auxiliary recognition factor cannot meet the preset threshold, establishing a continuous observation space according to the auxiliary recognition factor;
executing continuous data monitoring by initialized computing nodes in the continuous observation space;
and finishing data flow supervision and early warning according to the continuous data monitoring result.
According to experience, industry standards or related regulations, business requirements, risk preferences and actual conditions are comprehensively considered, and a proper preset threshold is formulated for determining the effectiveness of the auxiliary recognition factors. The auxiliary identification factor is compared with a preset threshold value, and if the factor exceeds the preset threshold value, the factor can be considered to meet the preset threshold value condition.
If the auxiliary identification factor cannot meet the preset threshold, the auxiliary identification factor does not trigger an alarm or supervision early warning of an abnormal condition. And reevaluating the auxiliary recognition factors which cannot meet the preset threshold, determining indexes suitable for establishing a continuous observation space based on the reevaluation result, wherein the indexes are related to abnormal behaviors or risks in financing service data, and can provide timely monitoring and warning functions. For each continuous observation index, proper thresholds and alarm rules are set, and the thresholds can be adjusted based on historical data, business requirements, risk preferences and other factors so as to capture abnormal conditions and trigger alarms in time. And establishing a continuous observation space according to the selected continuous observation index and the threshold, wherein the space comprises real-time data acquisition, automatic analysis, an alarm mechanism and the like, so that the system can continuously monitor and identify abnormal conditions related to auxiliary identification factors.
The continuous data source to be monitored is connected with the computing node, and continuous data monitoring is realized on the computing node according to the continuous observation space, including real-time statistical index calculation, anomaly detection and the like, so as to identify potential anomaly situations or anomaly modes.
Based on the results of the continuous data monitoring, an alarm or notification mechanism is triggered, such as sending a notification message, generating an alarm log to handle the abnormal situation, and at the same time, feeding the monitoring results back to the relevant personnel so that they can take necessary measures in time.
By performing continuous data monitoring on initialized computing nodes, the data in the continuous observation space can be monitored and analyzed in real time, any abnormal situation can be responded quickly, and the method is beneficial to improving the efficiency and accuracy of service operation and reducing risks and losses.
Further, the method further comprises:
establishing an evaluation mechanism of the computing node, wherein the evaluation mechanism comprises a speed evaluation mechanism and an accuracy evaluation mechanism;
and analyzing the processing effect of the node on the computing node through the evaluation mechanism, and updating the computing node based on the analysis result.
The speed evaluation mechanism comprises response time, throughput and expansibility, wherein the response time is used for measuring the time of a computing node responding to a request and comprises the time of data acquisition, processing and output, and the shorter response time is considered as an index of good performance; the throughput is used for measuring the number of requests which can be processed by the computing node in unit time, and the high throughput indicates that the computing node has better concurrent processing capacity; extensibility is used to evaluate the performance of a compute node as the load increases, which can test the computation by simulating and gradually increasing the load.
The accuracy evaluation mechanism comprises comparison verification, error rate analysis and data quality monitoring, wherein the comparison verification is used for comparing the result of the computing node with known accurate data or other verification methods, for example, a test data set of known answers is used for verifying whether the computing result of the computing node is consistent; error rate analysis is used to record the type and frequency of errors occurring in the compute nodes and conduct root cause analysis, which helps to determine key problems that improve the accuracy of the compute nodes and take corresponding measures to solve them; data quality monitoring is used to implement data quality monitoring measures, which help to improve the accuracy of the computing nodes and reduce errors due to low quality data.
And collecting performance and accuracy data of the computing nodes by using indexes defined in an evaluation mechanism, including collecting data such as response time, throughput, error rate and the like, analyzing the collected data, evaluating the processing effect of the computing nodes, for example, comparing the data under different time periods or different load conditions, and searching for anomalies.
Based on the analysis results, aspects of the computing node that require improvement are identified, which may include performance bottlenecks, accuracy defects, data quality problems, etc., in order to incorporate them into an improvement program. Based on the analysis result and the improvement plan, updating the computing node, and by means of optimization algorithm, hardware resource addition, parallel processing introduction and other methods, improving the response time, throughput and other performance indexes of the computing node to realize performance optimization; the architecture and configuration of the computing nodes are adjusted by increasing the number of nodes, reallocating workload, adjusting network topology, and the like, according to the analysis results and the improvement requirements.
After updating the compute nodes, using the test data sets, it is verified whether the processing effect of the nodes is improved and it is ensured that no new problems are introduced. This helps ensure that the compute nodes can meet changing demands and continue to provide high quality processing services.
In summary, the data flow supervision and early warning method and system for financing service provided by the embodiment of the application have the following technical effects:
1. by establishing a general data set and acquiring related data from a bank database, the integration of cross-bank and company main body data is realized, and a comprehensive financing business supervision foundation is provided;
2. by configuring the edge computing nodes and the matching network, real-time circulation data can be transmitted and processed more quickly, and real-time monitoring and analysis of data circulation are realized;
3. by constructing the node selection deflection factors and combining the feature set and the general data set of the real-time circulation data, the method can more accurately select proper computing nodes for data analysis, and improves the efficiency and accuracy;
4. comprehensive data flow supervision and early warning of financing business are realized through the combination of the first analysis result and the second analysis result, and potential risks and problems are effectively identified.
Therefore, the method achieves the technical effect of improving the supervision effect of the financing service data flow through solving the steps of data set integration, real-time processing, computing node selection, comprehensive analysis and the like.
Example two
Based on the same inventive concept as the data flow supervision and early warning method for financing service in the foregoing embodiment, as shown in fig. 2, the present application provides a data flow supervision and early warning system for financing service, the system includes:
the general data set establishing module 10 is used for establishing a general data set, the general data set is obtained through an interactive bank database, and after communication connection with the bank database is established, company main body data and bank main body data of financing service are read for construction;
the matching network generation module 20 is configured to configure a node update database of the edge computing node, and generate an edge computing node matching network by using the node update database as matching data;
the deviation factor constructing module 30 is used for obtaining the data flow supervision record of the current financing service, extracting the data flow supervision record and constructing a node selection deviation factor based on the extraction result;
The computing node selection module 40 is configured to read real-time circulation data, generate a data feature set of the real-time circulation data, and transmit the data feature set, the node selection bias factor and the general data set as configuration data to the edge computing node matching network to complete the selection and initialization of computing nodes;
the real-time data analysis module 50 is configured to execute data analysis corresponding to real-time circulation data through the initialized computing node, and generate a first analysis result;
the key data analysis module 60 is configured to extract key monitoring data of the computing nodes, send the key monitoring data to a processing center, and perform key monitoring data analysis of all computing nodes through the processing center to generate a second analysis result;
the data flow early warning module 70 is configured to perform data flow supervision early warning on the financing service according to the first analysis result and the second analysis result by the data flow early warning module 70.
Further, the system further comprises a first analysis result acquisition module for executing the following operation steps:
Extracting data position features in the data feature set, and sending the data position features to the position matching sub-network;
performing position matching calculation of edge calculation nodes on the data position features through the position matching sub-network to generate a position matching result;
extracting service characteristics of the data characteristic set through the service adaptation sub-network, and carrying out weight service adaptation matching on the extraction result and the general data set to generate a service matching result;
and sending the position matching result and the service matching result to the integration sub-network, executing normalization processing, and obtaining the first analysis result according to the processing result.
Further, the system also comprises a normalization processing module for executing the following operation steps:
performing network parameter compensation on the edge computing node matching network based on the node selection bias factor;
after the position matching result and the service matching result are sent to the integration sub-network, performing adaptive reconstruction of the position matching result and the service matching result through a network parameter compensation result;
and (5) finishing normalization processing by adapting the reconstruction result.
Further, the system also comprises a key monitoring data generation module for executing the following operation steps:
establishing an interest feature set of each data flow node, wherein the interest feature set is obtained by analyzing financing service;
when data analysis of corresponding real-time circulation data is carried out through a computing node, data extraction in a preset period is carried out through the interest feature set, and the data is recorded as a first data set;
if the data extraction result in any preset period is a null result, reselecting the mean characteristic data in the preset period, and marking the selection result as a second data set;
the critical monitoring data is generated from the first data set and the second data set.
Further, the system further comprises a data flow supervision and early warning module for executing the following operation steps:
carrying out overall anomaly analysis through the second analysis result, determining overall anomaly trend, and generating auxiliary recognition factors for local monitoring;
analyzing the first analysis result to obtain a local abnormal evaluation result, and optimizing a matching result of the local abnormal evaluation result through the auxiliary recognition factor;
and finishing the data flow supervision and early warning of the financing service through the optimized result.
Further, the system also comprises a supervision and early warning module for executing the following operation steps:
judging whether the auxiliary recognition factor meets a preset threshold value or not;
if the auxiliary recognition factor cannot meet the preset threshold, establishing a continuous observation space according to the auxiliary recognition factor;
executing continuous data monitoring by initialized computing nodes in the continuous observation space;
and finishing data flow supervision and early warning according to the continuous data monitoring result.
Further, the system also includes a computing node update module to perform the following operational steps:
establishing an evaluation mechanism of the computing node, wherein the evaluation mechanism comprises a speed evaluation mechanism and an accuracy evaluation mechanism;
and analyzing the processing effect of the node on the computing node through the evaluation mechanism, and updating the computing node based on the analysis result.
In the present disclosure, through the foregoing detailed description of the data flow supervision and early warning method for the financing service, those skilled in the art can clearly know the data flow supervision and early warning method and system for the financing service in this embodiment, and for the apparatus disclosed in the embodiment, the description is relatively simple because it corresponds to the method disclosed in the embodiment, and relevant places refer to the method part for description.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.