Business early warning management method and early warning management system based on data center

Abstract

The present invention belongs to the field of task management and data processing technology, and in particular, to a business warning management method and warning management system based on a data middle station. The method of the present invention is based on the powerful analysis capability of the data middle station, and by constructing various business data analysis models to perform smoothing / fitting processing and multi-dimensional comparative analysis on business data, abnormal data such as business data peaks are converted into business data deviations mined by the business data analysis model, thereby reducing false alarms caused by various data peaks. At the same time, the present invention establishes various dynamic time analysis models at the application layer, and the dynamic time analysis model performs secondary intervention on the business deviation data mined by the above-mentioned business data analysis model. The intervention process includes automatically adjusting the business warning threshold, thereby reducing the false alarm / missing phenomenon caused by the solidification of the business warning threshold, which can greatly improve the accuracy and timeliness of business warnings.

Description

Technical Field

[0001] The present invention belongs to the technical field of task management and data processing, and in particular, relates to a business early warning management method and an early warning management system based on a data center. Background Art

[0002] Nowadays, in order to discover operational problems as early as possible and reduce corporate losses, operating companies usually introduce business warning mechanisms during the digital transformation process. The current business warning mechanisms usually simply compare the business data analysis results with the preset business warning thresholds to determine whether to trigger a business warning. However, in this operating mode, there are the following problems:

[0003] (1) Since business data itself has various fluctuations, there are various data peaks in the actual operation process, which can easily lead to false alarms.

[0004] (2) The business warning threshold is fixed. Usually, a fixed business warning threshold is set based on the business experience of the business personnel. However, as the business scale changes, the fixed business warning threshold often deviates from the actual situation, resulting in abnormal situations such as false alarms or missed alarms. Summary of the invention

[0005] In order to overcome the above defects and deficiencies of the existing business warning mechanism, the present invention provides a set of solutions. The technical solution of the present invention adopts multi-level technical means to correct the above problems existing in the existing business warning mechanism. For problem (1), based on the powerful analysis ability of the data center, the present invention constructs various business data analysis models to smooth / fit the business data and conduct multi-dimensional comparative analysis, and converts abnormal data such as business data peaks into business data deviations mined by the business data analysis model, thereby reducing the false alarms caused by various data peaks; for problem (2), the present invention analyzes various historical data (including data of the enterprise, benchmark enterprises, and the industry) through business analysts, thereby establishing various dynamic time analysis models of the application layer, and for the business deviation data mined by the above business data analysis model (each enterprise can define its own application layer according to its own business characteristics), the dynamic time analysis model performs secondary intervention on it, and the intervention process includes automatically adjusting the business warning threshold, thereby reducing the false alarm / missing phenomenon caused by the solidification of the business warning threshold, and truly providing support for enterprise operations.

[0006] Based on the above design ideas, the present invention provides a business early warning management method and early warning management system that comprehensively applies multi-dimensional business data analysis based on the data middle station and business rule judgment technology based on the dynamic time analysis model. The method of the present invention is based on multiple data analysis results of the data middle station, combined with the dynamic business rule judgment mechanism of the application layer, which can greatly improve the accuracy and timeliness of business early warnings.

[0007] In the first aspect, the present invention provides a business early warning management method based on a data middle platform. The method configures business data processing rules, application layer rules and business data collection rules in the early warning management node, wherein the application layer rules include a dynamic time analysis model, and then guides the data collection node to complete the corresponding business data collection and submit the collected business data to the data middle platform; the data processing node located in the data middle platform performs a multi-dimensional analysis on the received business data according to the business data analysis model selected by the user submitted by the early warning management node or the business data analysis model automatically matched by the early warning management node according to the enterprise business model, and submits various business deviation data found during the analysis process to the early warning management node; the early warning management node adjusts the business early warning threshold according to the dynamic time analysis model and determines whether the business deviation data from the data middle platform triggers the business early warning process; the task management node controls the entire early warning process and supports multiple task management mechanisms, wherein the multiple task management mechanisms include timer tasks, manual scheduling triggered by management personnel, real-time tasks triggered by data streams provided by data collection nodes, and secondary analysis tasks triggered by business deviation data.

[0008] Furthermore, the service early warning management method of the present invention comprises the following steps:

[0009] Step 1: The early warning management node stores predefined business data processing rules, application layer rules and business data collection rules, wherein the application layer rules include various dynamic time analysis models;

[0010] Step 2: The user selects a business data analysis model, configures a data plan, a dynamic time analysis model, and a warning task handler, builds an overall early warning control plan, and submits the plan to the early warning management node;

[0011] Step 3: The early warning management node submits the data collection configuration to the data collection node, then starts the data collection node to collect business data and submits the collected business data to the data center;

[0012] Step 4: The early warning management node submits the data center analysis configuration to the data center, and then starts the data processing node located in the data center to perform business data analysis;

[0013] Step 5: The data center performs multi-dimensional analysis on the received business data. If an anomaly is found, it triggers internal processing of business deviations and submits the business deviation data and related environmental data to the early warning management node;

[0014] Step 6: The early warning management node determines whether the business deviation can constitute a warning trigger condition based on the dynamic time analysis model in the overall early warning control solution. If the trigger condition is met, the early warning task flow is constructed;

[0015] Step 7: The warning management node submits the warning information to the task management node. The task management node confirms the subsequent processing tasks and responsible persons according to the warning processing business flow rules, and returns the confirmation results to the warning management node;

[0016] Step 8: The early warning management node pushes the warning information and related task flow adaptation to the user's data terminal.

[0017] Furthermore, step 1 of the service early warning management method of the present invention specifically includes the following contents:

[0018] (1) The predefined business data processing rules stored in the early warning management node include the business data analysis model definition and business deviation definition used by the data center. The data center completes the abstract analysis of business data through the three-layer structure of business data analysis model, indicator system, and indicator analysis model. The business deviation data found by the indicator analysis model is the key point of the data layer for business early warning.

[0019] (2) The predefined application layer rules stored in the early warning management node include various dynamic time analysis models, early warning secondary judgment logic and empirical rules. The application layer rules are determined by the user according to the business model and business needs of the enterprise;

[0020] (3) The predefined business data collection rules stored in the early warning management node include the application rules of the entire data extraction-transformation-loading (ETL) process, including data source definition, data field definition, extraction rules, and collection time interval.

[0021] Furthermore, step 2 of the service early warning management method of the present invention specifically includes the following contents:

[0022] The user first confirms the core business data that needs to be analyzed based on the business model and business needs of the enterprise, and then performs relevant configurations around the core business data that needs to be analyzed through various data terminals. The configurations include the following categories:

[0023] (1) Data collection configuration: Based on the definition of the data warehouse ETL process, users configure relevant data sources, field definitions, extraction, planning, and storage models or parameters. During the data collection process, all raw data, intermediate data, and normalized result data are stored in the data lake located in the data center.

[0024] (2) Data center analysis configuration: Based on the core business data to be analyzed, the user configures the business data analysis model and screens the indicator system of interest, configures the indicator analysis model and related thresholds applied to business deviations, builds the business data analysis process and business deviation definitions, and the data center analyzes the various business data submitted by the data collection nodes according to the above configurations;

[0025] (3) Application layer rule configuration: Users select relevant dynamic time analysis models based on their business models and configure model parameters based on the time dimension. When analyzing business deviation data discovered by the data center, the dynamic time analysis model comprehensively utilizes model parameters of various time dimensions for unified judgment.

[0026] (4) Allocation of personnel to handle warning tasks: Allocation of personnel to handle warning tasks to ensure the timeliness of warning information processing.

[0027] Furthermore, step three of the service early warning management method of the present invention specifically includes the following contents:

[0028] The early warning management node submits the data collection configuration provided by the user to the data collection node, and then starts the data collection node to collect business data. For different business data sources, the data collection node has built-in data collection technology solutions that match them. The data collection technology solutions include data synchronization of various databases, reading of text files and field content extraction, and data interfaces based on http specifications.

[0029] According to the data collection configuration, the data collection node first synchronizes or builds the data dictionary definition, and then extracts and normalizes the business data fields from the original data source according to the data dictionary definition;

[0030] All raw data, intermediate data, and normalized result data processed by the data collection node are stored in the data lake located in the data center for reuse in subsequent secondary analysis and other business analysis.

[0031] Furthermore, step 4 of the service early warning management method of the present invention specifically includes the following contents:

[0032] (1) After the early warning management node submits the data center analysis configuration to the data center, the data center allocates computing resources to the corresponding business data analysis model and indicator analysis model according to the configuration, and then starts the data processing node located in the data center to perform business data analysis;

[0033] (2) The data processing nodes obtain various business data from the data lake. During the data analysis process, the data accessed by the data processing nodes include the customer's core business data and non-customer data selected according to the internal logic of the business data analysis model and the indicator analysis model.

[0034] Furthermore, step 5 of the service early warning management method of the present invention specifically includes the following contents:

[0035] (1) The data center is based on a three-layer structure of business data analysis model, indicator system, and indicator analysis model. According to the data center analysis configuration provided by the user, it dispatches relevant business data analysis models and indicator analysis models to process the customer's core business data. If a deviation in business indicators is found during the processing, the internal processing of the business deviation is immediately triggered;

[0036] (2) Internal processing of business deviations First, the environmental data that causes the business deviation is collected. The environmental data includes business data provided by the customer and external data. The data types included in the environmental data are pre-defined in the business deviation. The data center collects the corresponding environmental data based on this definition and submits it to the early warning management node.

[0037] Furthermore, step six of the service early warning management method of the present invention specifically includes the following contents:

[0038] (1) The dynamic time analysis model analyzes the business deviation data and environmental data, and combines the empirical criteria to determine whether the business deviation is upgraded to a warning sign; the dynamic time analysis model is a multivariate judgment model, and the variables include business indicator weights, business deviation amplitude, time and other auxiliary variables. The specific number of variables and parameter settings are comprehensively determined by the industry category, business model and channel type to which the business belongs;

[0039] (2) The early warning management node uses the secondary verification mechanism at the application layer to convert the business deviations found by the data center into warning information of different levels, and submits the warning information to the task management node. The task management node performs subsequent processing according to the enterprise's task management process.

[0040] Furthermore, step seven of the service early warning management method of the present invention specifically includes the following contents:

[0041] (1) The early warning management node and the data center cooperate to use a secondary verification mechanism to discover various warning signs. The early warning management node submits the warning information and related business data to the task management node, and the task management node starts the warning task processing mechanism;

[0042] (2) The task management node is responsible for maintaining the organization system and transmission system related to warning signs within the enterprise, and setting up different warning sign transmission systems and processing systems for warning signs of different business indicators and warning signs of different levels. In the initial stage, the user first builds the corresponding organization system and transmission system in the task management node according to different business indicators and warning sign levels, thereby forming a warning sign processing business flow rule base. After receiving the warning sign information submitted by the early warning management node, the task management node builds the warning sign processing business flow according to the rules defined in the warning sign processing business flow rule base, including confirming the relevant handlers, relevant informants and subsequent processing tasks of the warning sign.

[0043] (3) After the task management node constructs the warning processing business flow, it returns it to the warning management node;

[0044] Furthermore, step eight of the service early warning management method of the present invention specifically includes the following contents:

[0045] When the warning management node receives the warning processing business flow returned by the task management node, it means that the warning identification and warning processing process have been completed; at this time, the warning management node internally maintains the business deviation data, business deviation environment data, warning type, warning level, and warning processing process data of this warning; then the warning management node constructs the business deviation display chart, warning chart and task flow chart based on the above data, and pushes them to the data terminals of the team members involved in the task flow, so as to transfer to offline processing.

[0046] In a second aspect, the present invention provides a business warning management system based on a data middle platform, the system is composed of a data collection node, a warning management node, a warning rule library, a data middle platform, a task management node, a warning sign processing business flow rule library and an automatic control module, and the data middle platform includes a data processing node, a data lake, an analysis engine and a judgment library; wherein:

[0047] The function of the data collection node is to collect business data according to the data plan;

[0048] The functions of the early warning management node are to manage the early warning rule base and dynamic time analysis model, parse the overall early warning control solution pushed by the data terminal, control the data collection node to collect business data, judge the business deviation found by the data center through the application layer secondary verification mechanism, and output warning information and related task flows according to the member authority visualization adaptation;

[0049] The function of the warning rule library is to store warning rule information;

[0050] The function of the task management node is to manage the warning sign processing business flow rules and construct the warning sign processing business flow according to the warning sign processing business flow rule base;

[0051] The function of the warning sign processing business flow rule base is to store warning sign processing business flow rule information;

[0052] The function of the data processing node is to perform multi-dimensional analysis on business data based on the business model of the enterprise, and submit the business deviation data and related environmental data appearing in the analysis results to the early warning management node;

[0053] The function of the data lake is to store all raw data, intermediate data, and normalized result data collected and processed by the data collection nodes;

[0054] The function of the analysis engine is to cooperate with the data processing node to analyze the business data;

[0055] The function of the determination library is to store determination rule information;

[0056] The above-mentioned components operate according to the business early warning management method of the present invention, and the automatic control module automatically organizes and controls the components.

[0057] In summary, the method of the present invention is based on the powerful analysis capability of the data middle platform. By constructing various business data analysis models, the business data is smoothed / fitted and multi-dimensionally compared and analyzed. Abnormal data such as business data peaks are converted into business data deviations mined by the business data analysis model, thereby reducing the false alarms caused by various data peaks. At the same time, the present invention establishes various dynamic time analysis models at the application layer. The dynamic time analysis model performs secondary intervention on the business deviation data mined by the above-mentioned business data analysis model. The intervention process includes automatically adjusting the business warning threshold, thereby reducing the false alarm / omission caused by the solidification of the business warning threshold, which can greatly improve the accuracy and timeliness of business warnings. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention, not all embodiments. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0059] Figure 1 It is a schematic diagram of the implementation process of the method of the present invention.

[0060] Figure 2 Schematic diagram of the organizational structure of the method of the present invention. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. The present invention can also be implemented or applied through other different specific implementation methods. The details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention.

[0062] At the same time, it should be understood that the protection scope of the present invention is not limited to the specific embodiments described below; it should also be understood that the terms used in the embodiments of the present invention are for describing specific embodiments rather than for limiting the protection scope of the present invention.

[0063] Example 1: A business warning management method based on a data center (see Figure 1-2 ), the method configures business data processing rules, application layer rules and business data collection rules in the early warning management node, wherein the application layer rules include a dynamic time analysis model, and then guides the data collection node to complete the corresponding business data collection and submit the collected business data to the data middle station; the data processing node located in the data middle station performs multi-dimensional analysis on the received business data according to the business data analysis model selected by the user submitted by the early warning management node or the business data analysis model automatically matched by the early warning management node according to the enterprise business model, and submits various business deviation data found in the analysis process to the early warning management node; the early warning management node adjusts the business early warning threshold according to the dynamic time analysis model and determines whether the business deviation data from the data middle station triggers the business early warning process; the task management node controls the entire early warning process and supports multiple task management mechanisms, wherein the multiple task management mechanisms include timer tasks, manual scheduling triggered by management personnel, real-time tasks triggered by data streams provided by data collection nodes, and secondary analysis tasks triggered by business deviation data.

[0064] Specifically, the method comprises the following steps:

[0065] Step 1: The early warning management node stores predefined business data processing rules, application layer rules and business data collection rules, wherein the application layer rules include various dynamic time analysis models.

[0066] (1) The predefined business data processing rules stored in the early warning management node include the business data analysis model definition and business deviation definition used by the data center. The data center has accumulated various business data analysis model definitions and business deviation definitions through long-term operation. For example, for sales flow data, the data center provides commodity value analysis models, price band analysis models, brand benefit analysis models, region / store analysis models, channel / platform analysis models, RFM models, etc., and based on these analysis models, a complete set of sales indicator systems is constructed, including sales, order volume, completion rate, linkage rate, growth rate, sales share of key commodities, sales share of each platform, etc., and then the business deviations existing in these indicators are mined through the regular analysis model / comparative analysis model of business indicators. Finally, the data center completes the abstract analysis of business data through the three-layer structure of business data analysis model-indicator system-indicator analysis model, among which the business deviation data discovered by the indicator analysis model is the key point of the data layer of business early warning.

[0067] (2) The predefined application layer rules stored in the early warning management node include various dynamic time analysis models, early warning secondary judgment logic (for example, three consecutive data layer warnings within 10 minutes can be judged as real business warnings) and empirical rules. The above application layer rules are determined by users based on the business model and business needs of the enterprise. For example, for sales volume indicators, there are large differences in the definition of dynamic time analysis models in different industries / enterprises. For example, when analyzing the sales volume indicators of clothing sales companies, each quarter [off-season / peak season], each month, and each week [working day / rest day] have different model parameters and fitting curves; in this way, the sales volume deviation data found by the data center can be analyzed again by the dynamic time analysis model of the application layer before it can be included in the early warning rules as a warning sign for processing.

[0068] (3) The predefined business data collection rules stored in the early warning management node include the application rules of the entire data extraction-transformation-loading (ETL) process, including data source definition, data field definition, extraction rules, and collection time interval.

[0069] Step 2: The user selects a business data analysis model, configures the data plan, dynamic time analysis model and warning task handler, builds an overall early warning control plan, and submits the plan to the early warning management node.

[0070] The user first confirms the core business data that needs to be analyzed based on the business model and business needs of the enterprise, such as sales data, financial data, personnel performance data, etc. Then, the user performs relevant configurations around the core business data that needs to be analyzed through various data terminals (mobile phones, PCs, and the Web). The configurations include the following categories:

[0071] (1) Data collection configuration: Based on the definition of the data warehouse ETL process, users configure relevant data sources, field definitions, extraction, planning, and storage models or parameters. During the data collection process, all raw data, intermediate data, and normalized result data are stored in the data lake located in the data center.

[0072] (2) Data center analysis configuration: Based on the core business data to be analyzed, the user configures the business data analysis model and screens the indicator system of interest, configures the indicator analysis model and related thresholds applied to business deviations, builds the business data analysis process and business deviation definitions, and the data center analyzes the various business data submitted by the data collection nodes according to the above configurations;

[0073] (3) Application layer rule configuration: Users select relevant dynamic time analysis models based on the business model and configure model parameters based on the time dimension, such as model parameters used for each quarter, month, and week. When analyzing the business deviation data found by the data center, the dynamic time analysis model comprehensively utilizes the model parameters of various time dimensions for unified judgment. For example, for clothing sales companies, deviation fitting curves will be constructed based on model parameters at the quarterly / monthly / weekly levels according to different weights. Application layer rules also include commonly used empirical rules. For example, for sales day data analysis, business deviation data is usually upgraded to a warning sign and reported to business managers only when it is below the deviation fitting curve for three consecutive days.

[0074] (4) Alarm task processing personnel configuration: As a complete management system, alarm task processing personnel must be configured to ensure the timeliness of alarm information processing.

[0075] Step 3: The early warning management node submits the data collection configuration to the data collection node, then starts the data collection node to collect business data and submits the collected business data to the data center.

[0076] The early warning management node submits the data collection configuration provided by the user to the data collection node, and then starts the data collection node to collect business data. For different business data sources, the data collection node has built-in data collection technology solutions that match them. The data collection technology solutions include data synchronization of various databases, reading of text files and field content extraction, and data interfaces based on http specifications.

[0077] According to the data collection configuration, the data collection node first synchronizes or builds the data dictionary definition, and then extracts and normalizes the business data fields from the original data source according to the data dictionary definition. For example, time field normalization, currency field normalization, etc. For another example, for sales flow data, the constructed data dictionary defines business fields such as serial number, product SKU, product name, unit price, sales volume, sales date, sales time, channel ID, channel name, store ID, store name, etc., and defines the correspondence between the above business fields and various original data content fragments, extraction methods, conversion methods, etc.

[0078] All raw data, intermediate data, and normalized result data processed by the data collection node are stored in the data lake located in the data center for reuse in subsequent secondary analysis and other business analysis.

[0079] Step 4: The early warning management node submits the data center analysis configuration to the data center, and then starts the data processing node located in the data center to perform business data analysis.

[0080] (1) After the early warning management node submits the data center analysis configuration to the data center, the data center allocates computing resources to the corresponding business data analysis model and indicator analysis model according to the configuration, and then starts the data processing node located in the data center to perform business data analysis;

[0081] (2) Data processing nodes obtain various business data from the data lake. During the data analysis process, the data accessed by the data processing nodes include the customer's core business data and non-customer data selected according to the internal logic of the business data analysis model and the indicator analysis model. For example, in the sales flow data analysis process of a clothing retail enterprise, some industry data and public data of channels / platforms may also be used, such as competitor product prices and review data on the e-commerce platform.

[0082] Step 5: The data center performs a multi-dimensional analysis of the received business data. If an anomaly is found, it triggers internal processing of business deviations and submits the business deviation data and related environmental data to the early warning management node.

[0083] (1) The data center is based on the three-layer structure of business data analysis model-indicator system-indicator analysis model. According to the data center analysis configuration provided by the user, it dispatches the relevant business data analysis model and indicator analysis model to process the customer's core business data. If the business indicator is found to be deviated during the processing, the internal processing of the business deviation is triggered immediately. For example, in the sales flow data analysis process of clothing retail enterprises, when analyzing the daily sales flow of customers, the commodity value analysis model is usually configured to analyze the popularity of various types of clothing (expressed by indicators such as daily sales volume / daily sales revenue / daily linkage rate); regional / store analysis models are also used to analyze the differences between stores, etc.; if the user has defined the business deviation range for the "daily sales revenue" business indicator in the data center analysis configuration (such as less than 60% of the daily average of the previous ten days), it will trigger the internal processing of the business deviation if it is found during the analysis process that the daily sales revenue of the day triggers the deviation definition configured by the customer.

[0084] (2) Internal processing of business deviations First, environmental data that causes business deviations is collected. The environmental data includes business data (commodity price data, etc.) and external data (date / holiday data, store weather data, etc.) provided by customers. The types of data included in the environmental data are pre-defined in the business deviation. The data center collects the corresponding environmental data based on this definition and submits it to the early warning management node. For example, in the sales flow data analysis process of a clothing retail enterprise, if the deviation definition of the "daily sales" indicator is triggered, environmental data containing store weather, date / holiday and other external data will usually be collected according to the configuration.

[0085] Step 6: The early warning management node determines whether the business deviation can constitute a warning trigger condition based on the dynamic time analysis model in the overall early warning control plan. If the trigger condition is met, the early warning task flow is constructed.

[0086] (1) The dynamic time analysis model analyzes business deviation data and environmental data, and combines empirical criteria to determine whether the business deviation has been upgraded to a warning sign; the dynamic time analysis model is a multivariate judgment model, and the variables include business indicator weights, business deviation amplitude, time and other auxiliary variables. The specific number of variables and parameter settings are determined by the industry category, business model and channel type to which the business belongs. For example, for the sales flow data analysis of clothing retail enterprises, for the regional / store business model, the spatial dimension information such as the weather in the area where the store is located and the per capita income in the area where the store is located is usually included in the model as a variable; for the e-commerce business model, the market share of each e-commerce platform and the additional service evaluation value of the e-commerce platform (such as express service, after-sales support) are usually included in the model as variables.

[0087] (2) The early warning management node uses the secondary verification mechanism at the application layer to convert the business deviations found by the data center into warning information of different levels, and submits the warning information to the task management node. The task management node performs subsequent processing according to the enterprise's task management process.

[0088] Dynamic Time Analysis Model Example

[0089] For example, for clothing retailers that adopt a regional / store business model, the following variables are usually incorporated into their dynamic time analysis model to address business deviations caused by daily sales indicators.

[0090] X1--Daily sales weight;

[0091] X2--Daily sales business deviation value;

[0092] X3--holiday / week of business date;

[0093] X4--Store weather;

[0094] X5--the per capita income of the area where the store is located (last year's value);

[0095] X6--Estimated population of the area where the store is located;

[0096] …

[0097] These variables construct the multivariate decision function:

[0098] w=F(X1,X2,X3,X4,X5,X6…);

[0099] The multivariate decision function is used to perform a secondary judgment on the business deviation caused by the daily sales index of the clothing retail enterprise, and the result w is a value in the interval (0, 1), which is used to indicate the probability of the business deviation being upgraded to a warning sign.

[0100] By introducing multiple variables according to the above method, a single business deviation can be evaluated in multiple dimensions. For example, for the same business deviation of a 60% drop in daily sales, if the weather in the store area on that day is extreme weather or it is the first working day after the National Day holiday, the result value of the multivariate judgment function may be relatively low, thereby avoiding the upgrade of the business deviation to a warning sign.

[0101] The application layer rule configuration also includes commonly used empirical criteria, which are applied to multiple aspects. For example, for the result value w of the multivariate decision function, the empirical criteria are usually used to classify w as follows:

[0102] 1>w>=0.8 Red warning sign;

[0103] 0.8>w>=0.5 Yellow warning sign;

[0104] 0.5>w Green (normal).

[0105] Users can also adopt other classifications for w according to their company's early warning management system and business needs.

[0106] The application layer rule configuration adopts a mechanism that combines empirical methods and historical experience methods. To this end, at the beginning of the application, users need to use historical data to adjust the application layer rule configuration multiple times, evaluate the results of each time, and determine the model parameters and experience rules that are most suitable for the enterprise. In the subsequent operation process, users also need to track the results and fine-tune the model parameters.

[0107] Step 7: The warning management node submits the warning information to the task management node. The task management node confirms the subsequent processing tasks and responsible persons according to the warning processing business flow rules, and returns the confirmation results to the warning management node.

[0108] (1) The early warning management node and the data center cooperate to use a secondary verification mechanism to discover various warning signs. The early warning management node submits the warning information and related business data to the task management node, and the task management node starts the warning task processing mechanism;

[0109] (2) The task management node is responsible for maintaining the organization system and transmission system related to warning signs within the enterprise, and setting up different warning sign transmission systems and processing systems for warning signs of different business indicators and warning signs of different levels (such as red / yellow). In the initial stage, the user first builds the corresponding organization system and transmission system in the task management node according to different business indicators and warning sign levels, thereby forming a warning sign processing business flow rule base. After receiving the warning sign information submitted by the early warning management node, the task management node builds the warning sign processing business flow according to the rules defined in the warning sign processing business flow rule base, including confirming the relevant handlers, relevant informants and subsequent processing tasks of the warning sign.

[0110] (3) After the task management node constructs the warning processing business flow, it returns it to the warning management node.

[0111] Step 8: The early warning management node pushes the warning information and related task flow adaptation to the user's data terminal.

[0112] When the warning management node receives the warning processing business flow returned by the task management node, it means that the warning identification and warning processing process have been completed; at this time, the warning management node internally maintains the business deviation data, business deviation environment data, warning type, warning level, and warning processing process data of this warning; then the warning management node constructs the business deviation display chart, warning chart and task flow chart based on the above data, and pushes them to the data terminals of the team members involved in the task flow, so as to transfer to offline processing.

[0113] Example 2: A business warning management system based on a data center (see Figure 2 )

[0114] The system consists of a data collection node, an early warning management node, an early warning rule library, a data center, a task management node, an early warning business flow rule library, and an automatic control module. The data center includes a data processing node, a data lake, an analysis engine, and a judgment library; wherein:

[0115] The function of the data collection node is to collect business data according to the data plan;

[0116] The functions of the early warning management node are to manage the early warning rule base and dynamic time analysis model, parse the overall early warning control solution pushed by the data terminal, control the data collection node to collect business data, judge the business deviation found by the data center through the application layer secondary verification mechanism, and output warning information and related task flows according to the member authority visualization adaptation;

[0117] The function of the warning rule library is to store warning rule information;

[0118] The function of the task management node is to manage the warning sign processing business flow rules and construct the warning sign processing business flow according to the warning sign processing business flow rule base;

[0119] The function of the warning sign processing business flow rule base is to store warning sign processing business flow rule information;

[0120] The function of the data processing node is to perform multi-dimensional analysis on business data based on the business model of the enterprise, and submit the business deviation data and related environmental data appearing in the analysis results to the early warning management node;

[0121] The function of the data lake is to store all raw data, intermediate data, and normalized result data collected and processed by the data collection nodes;

[0122] The function of the analysis engine is to cooperate with the data processing node to analyze the business data;

[0123] The function of the determination library is to store determination rule information;

[0124] The above-mentioned components operate according to the business early warning management method of Example 1, and the automatic control module automatically organizes and controls the components.

[0125] Embodiment 3: An electronic device for business warning management

[0126] The electronic device includes a processor and a memory, the memory is used to store programs, and the processor is used to run the programs to implement the business early warning management method based on the data middle station described in Example 1.

[0127] The various embodiments of the present invention are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0128] The above description is only an embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, replacement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the claims of the present invention.

Claims

1. A business early warning management method based on a data center, characterized in that: This method configures business data processing rules, application layer rules and business data collection rules in the early warning management node, wherein the application layer rules include a dynamic time analysis model, and then guides the data collection node to complete the corresponding business data collection and submit the collected business data to the data middle station; The data processing node located in the data center performs multi-dimensional analysis on the received business data according to the business data analysis model selected by the user submitted by the early warning management node or the business data analysis model automatically matched by the early warning management node according to the enterprise business model, and submits various business deviation data found during the analysis process to the early warning management node; The warning management node adjusts the business warning threshold according to the dynamic time analysis model and determines whether the business deviation data from the data center triggers the business warning process; The task management node controls the entire warning process and supports multiple task management mechanisms, including timer tasks, manual scheduling triggered by management personnel, real-time tasks triggered by data streams provided by data collection nodes, and secondary analysis tasks triggered by business deviation data; The method comprises the following steps: Step 1: The early warning management node stores predefined business data processing rules, application layer rules and business data collection rules, wherein the application layer rules include various dynamic time analysis models; Step 2: The user selects a business data analysis model, configures a data plan, a dynamic time analysis model, and a warning task handler, builds an overall early warning control plan, and submits the plan to the early warning management node; Step 3: The early warning management node submits the data collection configuration to the data collection node, then starts the data collection node to collect business data and submits the collected business data to the data center; Step 4: The early warning management node submits the data center analysis configuration to the data center, and then starts the data processing node located in the data center to perform business data analysis; Step 5: The data center performs multi-dimensional analysis on the received business data. If an anomaly is found, it triggers internal processing of business deviations and submits the business deviation data and related environmental data to the early warning management node; Step 6: The early warning management node determines whether the current business deviation can constitute a warning trigger condition based on the dynamic time analysis model in the overall early warning control solution. If the trigger condition is met, the early warning task flow is constructed; Step 7: The warning management node submits the warning information to the task management node. The task management node confirms the subsequent processing tasks and responsible persons according to the warning processing business flow rules, and returns the confirmation results to the warning management node; Step 8: The early warning management node pushes the warning information and related task flow adaptation to the user's data terminal.

2. The business early warning management method according to claim 1, characterized in that: Step 1 specifically includes the following: (1) The predefined business data processing rules stored in the early warning management node include the business data analysis model definition and business deviation definition used by the data center. The data center completes the abstract analysis of business data through the three-layer structure of business data analysis model, indicator system, and indicator analysis model. The business deviation data found by the indicator analysis model is the key point of the data layer for business early warning. (2) The predefined application layer rules stored in the early warning management node include various dynamic time analysis models, early warning secondary judgment logic and empirical rules. The application layer rules are determined by the user according to the business model and business needs of the enterprise; (3) The predefined business data collection rules stored in the early warning management node include the application rules for the entire process of data extraction, conversion, and loading, including data source definition, data field definition, extraction rules, and collection time interval.

3. The business early warning management method according to claim 1, characterized in that: Step 2 specifically includes the following: The user first confirms the core business data that needs to be analyzed based on the business model and business needs of the enterprise, and then performs relevant configurations around the core business data that needs to be analyzed through various data terminals. The configurations include the following categories: (1) Data collection configuration: Based on the definition of the data warehouse ETL process, users configure relevant data sources, field definitions, extraction, planning, and storage models or parameters. During the data collection process, all original data, intermediate data, and normalized result data are stored in the data lake located in the data center; (2) Data center analysis configuration: Based on the core business data to be analyzed, the user configures the business data analysis model and screens the indicator system of interest, and configures the indicator analysis model and related thresholds applied to business deviations, builds the business data analysis process and business deviation definition. The data center analyzes the various business data submitted by the data collection node according to the above configuration; (3) Application layer rule configuration: Users select relevant dynamic time analysis models based on their business models and configure model parameters based on the time dimension. When analyzing business deviation data discovered by the data center, the dynamic time analysis model comprehensively utilizes model parameters of various time dimensions for unified judgment. (4) Allocation of personnel to handle warning tasks: Allocation of personnel to handle warning tasks to ensure the timeliness of warning information processing.

4. The business early warning management method according to claim 1, characterized in that: Step 3 specifically includes the following: The early warning management node submits the data collection configuration provided by the user to the data collection node, and then starts the data collection node to collect business data. For different business data sources, the data collection node has built-in data collection technology solutions that match them. The data collection technology solutions include data synchronization of various databases, reading of text files and field content extraction, and data interfaces based on http specifications. According to the data collection configuration, the data collection node first synchronizes or builds the data dictionary definition, and then extracts and normalizes the business data fields from the original data source according to the data dictionary definition; All raw data, intermediate data, and normalized result data processed by the data collection node are stored in the data lake located in the data center for reuse in subsequent secondary analysis and other business analysis.

5. The business early warning management method according to claim 1, characterized in that: Step 4 specifically includes the following: (1) After the early warning management node submits the data center analysis configuration to the data center, the data center allocates computing resources to the corresponding business data analysis model and indicator analysis model according to the configuration, and then starts the data processing node located in the data center to perform business data analysis; (2) The data processing nodes obtain various business data from the data lake. During the data analysis process, the data accessed by the data processing nodes include the customer's core business data and non-customer data selected according to the internal logic of the business data analysis model and the indicator analysis model.

6. The business early warning management method according to claim 1, characterized in that: Step 5 specifically includes the following: (1) The data center is based on the three-layer structure of business data analysis model, indicator system, and indicator analysis model. According to the data center analysis configuration provided by the user, it dispatches the relevant business data analysis model and indicator analysis model to process the customer's core business data. If a deviation in business indicators is found during the processing, the internal processing of the business deviation is immediately triggered; (2) Internal processing of business deviations First, the environmental data that causes the business deviation is collected. The environmental data includes business data provided by the customer and external data. The data types included in the environmental data are pre-defined in the business deviation. The data center collects the corresponding environmental data based on this definition and submits it to the early warning management node.

7. The business early warning management method according to claim 1, characterized in that: Step 6 specifically includes the following: (1) The dynamic time analysis model analyzes the business deviation data and environmental data, and combines the empirical criteria to determine whether the business deviation is upgraded to a warning sign; the dynamic time analysis model is a multivariate judgment model, and the variables include business indicator weights, business deviation amplitude, time and other auxiliary variables. The specific number of variables and parameter settings are comprehensively determined by the industry category, business model and channel type to which the business belongs; (2) The early warning management node uses the secondary verification mechanism at the application layer to ultimately convert the business deviations discovered by the data center into warning information of different levels, and submits the warning information to the task management node. The task management node then performs subsequent processing based on the enterprise's task management process.

8. The business early warning management method according to claim 1, characterized in that: Step 7 specifically includes the following: (1) The early warning management node and the data center work together to use a secondary verification mechanism to discover various warning signs. The early warning management node submits the warning information and related business data to the task management node, and the task management node starts the warning task processing mechanism; (2) The task management node is responsible for maintaining the internal organizational system and transmission system related to warning signs, and setting up different warning sign transmission systems and processing systems for warning signs of different business indicators and warning signs of different levels; In the initial stage, users first build corresponding organizational systems and transmission systems in the task management node according to different business indicators and warning levels, thus forming a warning processing business flow rule base; after receiving the warning information submitted by the early warning management node, the task management node builds the warning processing business flow according to the rules defined in the warning processing business flow rule base, including confirming the relevant processors, relevant informants and subsequent processing tasks of the warning; (3) After the task management node constructs the warning processing business flow, it returns it to the warning management node; Step 8 Includes the following: When the warning management node receives the warning processing business flow returned by the task management node, it means that the warning identification and warning processing process have been established; at this time, the warning management node internally maintains the business deviation data, business deviation environment data, warning type, warning level, and warning processing process data of this warning; The early warning management node then constructs business deviation display charts, warning charts, and task flow charts based on the above data, and pushes them to the data terminals of the team members involved in the task flow, thereby transferring them to offline processing.

9. A business early warning management system based on a data center, characterized in that: The system consists of a data collection node, an early warning management node, an early warning rule library, a data center, a task management node, an early warning business flow rule library, and an automatic control module. The data center includes a data processing node, a data lake, an analysis engine, and a judgment library; wherein: The function of the data collection node is to collect business data according to the data plan; The functions of the early warning management node are to manage the early warning rule base and dynamic time analysis model, parse the overall early warning control solution pushed by the data terminal, control the data collection node to collect business data, judge the business deviation found by the data center through the application layer secondary verification mechanism, and output warning information and related task flows according to the member authority visualization adaptation; The function of the warning rule library is to store warning rule information; The function of the task management node is to manage the warning sign processing business flow rules and construct the warning sign processing business flow according to the warning sign processing business flow rule base; The function of the warning sign processing business flow rule base is to store warning sign processing business flow rule information; The function of the data processing node is to perform multi-dimensional analysis on business data based on the business model of the enterprise, and submit the business deviation data and related environmental data appearing in the analysis results to the early warning management node; The function of the data lake is to store all raw data, intermediate data, and normalized result data collected and processed by the data collection nodes; The function of the analysis engine is to cooperate with the data processing node to analyze the business data; The function of the determination library is to store determination rule information; The system operates according to the business early warning management method described in any one of claims 1 to 8, and the automatic control module automatically organizes and controls each component.

Images