CN116955315A

CN116955315A - Migration model construction method, migration model construction device, migration model construction equipment and storage medium

Info

Publication number: CN116955315A
Application number: CN202310857520.3A
Authority: CN
Inventors: 敖倩; 黄钊港; 陈灿然; 孙海鑫; 徐一茗; 付小奇
Original assignee: Bank of China Ltd
Current assignee: Bank of China Ltd
Priority date: 2023-07-13
Filing date: 2023-07-13
Publication date: 2023-10-27

Abstract

The application discloses a method, a device, equipment and a storage medium for constructing a migration model, which are applied to the field of big data or the field of finance. When the method provided by the embodiment of the application is executed, the enterprise frame migration rule, the enterprise frame message processing rule and the old framework core field can be obtained first, and the blood margin analysis technology is utilized to carry out blood margin analysis on the enterprise frame migration rule, the enterprise frame message processing rule and the old framework core field to obtain a blood margin relation set. And performing migration learning on the blood relationship set by using a migration learning algorithm to obtain a preliminary migration model, and performing migration learning on the first message system by using the migration learning algorithm based on the preliminary migration model to obtain a modified migration model. And acquiring a unique theme and a shared theme of the second message system relative to the first message system by using a recommendation system algorithm, and supplementing the transformation migration model by using the second message system, the unique theme and the shared theme to obtain the migration model. The application improves the accuracy and the data quality of the reconstruction of the penholder.

Description

Migration model construction method, migration model construction device, migration model construction equipment and storage medium

Technical Field

The application relates to the field of big data operation and finance, in particular to a method, a device, equipment and a storage medium for constructing a migration model.

Background

Enterprise architecture construction refers to the process of planning, designing, and implementing an enterprise overall architecture within an organization. It aims to ensure that all aspects of the organization can work cooperatively and efficiently and achieve the strategic goals of the organization. The goal of enterprise architecture construction is to achieve a high degree of business and technology compliance, improving the overall efficiency and flexibility of the organization. Through planning and optimizing the enterprise architecture, organizations can better cope with market changes, promote business innovations, and provide high-quality products and services.

The influence scope of the new architecture on each topic message system is not equal under the construction of the enterprise architecture in the bank system, and the enterprise architecture migration group can migrate to the component data under the new architecture only according to the core data under the old architecture. The topic message systems have differences in component data analysis under a new architecture due to topic differences, rule differences and the like, so that the communication cost of a migration group and each project group is increased, the analysis cost of subsequent transformation of the project group is also increased, and the migration efficiency is reduced.

Therefore, how to improve the accuracy and data quality of the transformation of the enterprise shelf, so as to reduce the communication cost of the enterprise shelf migration group and each project group and the analysis cost of the subsequent transformation of the project group is a technical problem which needs to be solved by the technicians in the field.

Disclosure of Invention

Based on the problems, the application provides a construction method, a device, equipment and a storage medium of a migration model, which can improve the accuracy and the data quality of enterprise transformation, thereby reducing the communication cost of enterprise migration groups and project groups and the analysis cost of subsequent transformation of the project groups.

The embodiment of the application discloses the following technical scheme:

a method of constructing a migration model, the method comprising:

obtaining enterprise frame migration rules, enterprise frame message processing rules and old framework core fields, and performing blood margin analysis on the enterprise frame migration rules, the enterprise frame message processing rules and the old framework core fields by utilizing a blood margin analysis technology to obtain a blood margin relationship set;

performing migration learning on the blood relationship set by using a migration learning algorithm to obtain a preliminary migration model;

performing transfer learning on the first message system by using the transfer learning algorithm based on the preliminary transfer model to obtain a modified transfer model;

Acquiring a unique theme and a shared theme of a second message system relative to the first message system by using a recommendation system algorithm;

and supplementing the transformation migration model by using the second message system, the unique theme and the shared theme to obtain a migration model.

In one possible implementation manner, the performing the migration learning on the first message system by using the migration learning algorithm based on the preliminary migration model to obtain a modified migration model includes:

acquiring a message index processing rule, a message index field set and a blood-edge relation between a first message system and an old framework of the first message system;

inputting the message index processing rule, the message index field set and the blood relationship between the first message system and the old framework into the preliminary migration model, and performing migration learning by using the migration learning algorithm to obtain a migration model to be tuned;

and carrying out parameter fine adjustment on the migration model to be adjusted to obtain the transformation migration model.

In one possible implementation, the set of blood-lineage relationships includes: migrating old architecture blood-edge relationships, message old architecture blood-edge relationships, and migrating message blood-edge relationships.

In one possible implementation, the method further includes:

and storing the blood edge relation in the blood edge relation set in a paste source layer database.

In one possible implementation, the recommender system algorithm is a TPLSA-Imp algorithm.

In a possible implementation manner, the supplementing the modified migration model with the second messaging system, the unique topic, and the shared topic to obtain a migration model includes:

acquiring the blood-edge relation between the second message system and the old framework;

determining an index processing rule of the second message system by utilizing the blood-edge relationship between the second message system and the old framework;

extracting unique characteristic indexes of the unique subject and shared characteristic indexes of the shared subject through the index processing rules;

and supplementing the transformation migration model by utilizing the unique characteristic indexes and the shared characteristic indexes to obtain a migration model.

A migration model building apparatus, the apparatus comprising:

the acquisition and analysis unit is used for acquiring enterprise frame migration rules, enterprise frame message processing rules and old framework core fields, and performing blood margin analysis on the enterprise frame migration rules, the enterprise frame message processing rules and the old framework core fields by utilizing a blood margin analysis technology to obtain a blood margin relationship set;

The first learning unit is used for performing migration learning on the blood relationship set by using a migration learning algorithm to obtain a preliminary migration model;

the second learning unit is used for performing transfer learning on the first message system by utilizing the transfer learning algorithm based on the preliminary transfer model to obtain a modified transfer model;

the first acquisition unit is used for acquiring the unique theme and the shared theme of the second message system relative to the first message system by using a recommendation system algorithm;

the first supplementing unit is used for supplementing the transformation migration model by using the second message system, the unique theme and the shared theme to obtain a migration model.

In one possible implementation, the apparatus further includes:

the second acquisition unit is used for acquiring the message index processing rule, the message index field set and the blood-edge relationship between the first message system and the old framework of the first message system;

the input unit is used for inputting the message index processing rule, the message index field set, the first message system and the blood relationship of the old framework into the preliminary migration model, and performing migration learning by utilizing the migration learning algorithm to obtain a migration model to be tuned;

And the parameter adjustment unit is used for carrying out parameter fine adjustment on the migration model to be adjusted to obtain the transformation migration model.

A migration model building apparatus, comprising: the migration model constructing method comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, wherein the processor realizes the migration model constructing method when executing the computer program.

A computer readable storage medium having instructions stored therein which, when executed on a terminal device, cause the terminal device to perform a method of constructing a migration model as described above.

The application provides a migration model construction method, a migration model construction device, migration model construction equipment and a storage medium. Specifically, when the method for constructing the migration model provided by the embodiment of the application is executed, the enterprise migration rule, the enterprise message processing rule and the old architecture core field can be obtained first, and the blood margin analysis technology is utilized to carry out blood margin analysis on the enterprise migration rule, the enterprise message processing rule and the old architecture core field to obtain a blood margin relation set. And then, performing migration learning on the blood relationship set by using a migration learning algorithm to obtain a preliminary migration model, and performing migration learning on the first message system by using the migration learning algorithm based on the preliminary migration model to obtain a modified migration model. And then acquiring a unique theme and a shared theme of the second message system relative to the first message system by using a recommendation system algorithm, and supplementing the modified migration model by using the second message system, the unique theme and the shared theme to obtain the migration model. According to the application, the blood relationship among the enterprise shelf migration rules, the enterprise shelf message processing rules and the old architecture core fields is traced through the blood relationship analysis technology to obtain a blood relationship set, the blood relationship set and the first message system are subjected to migration learning by using a migration learning algorithm to obtain a modified migration model, and the second message system is analyzed based on a recommendation system algorithm to finally obtain a unique theme and a shared theme of the second message system which can be directly used for supplementing the modified migration model, so that the accuracy and the data quality of enterprise shelf modification can be improved by constructing the migration model, and the communication cost of an enterprise shelf migration group and each project group and the analysis cost of subsequent modification of the project group are reduced.

Drawings

In order to more clearly illustrate this embodiment or the technical solutions of the prior art, the drawings that are required for the description of the embodiment or the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a flow chart of a method for constructing a migration model according to an embodiment of the present application;

fig. 2 is a schematic structural diagram of a migration model building apparatus according to an embodiment of the present application.

Detailed Description

In order to make the present application better understood by those skilled in the art, the following description will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings, and it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.

In order to facilitate understanding of the technical solution provided by the embodiments of the present application, the following description will first explain the background technology related to the embodiments of the present application.

Enterprise architecture construction, among other things, generally includes the following aspects: (1) service architecture: the business objectives, business flows, and business rules of an organization, as well as relationships and dependencies between businesses, are defined. It helps the organization understand its own core traffic, providing the basis for other architecture hierarchies. (2) And (3) data architecture: data assets, data flows, and data management policies of an organization are defined. It ensures that data can provide support for decisions and operations of an organization, taking into account the integrity, availability and security of the data. (3) Application architecture: application systems, software and technology platforms of an organization are defined. It includes the integration between application systems, the functional and performance requirements of the application systems, and specifications for software development and maintenance. (4) The technical architecture is as follows: technical infrastructure defining an organization, such as a network, server, storage, etc. It provides reliable support for applications and data in view of the technical requirements and cost effectiveness of the organization.

An enterprise architecture migration group is a team responsible for organizing and managing the enterprise architecture migration process. They play a key role during enterprise architecture revolution, ensuring the smooth progress of the entire migration process. The main responsibilities of the enterprise architecture migration group include: (1) planning and designing: the enterprise architecture migration group is responsible for making migration plans and strategies, and designs new enterprise architecture schemes according to strategic goals and demands of organizations. They consider the architecture of the various layers, such as business, data, applications and technology, and ensure that the new architecture is consistent with the goals of the organization. (2) Evaluation and analysis: the enterprise architecture migration group evaluates and analyzes the existing enterprise architecture and knows the limitations and problems of the existing enterprise architecture. They identify areas that need improvement or optimization and propose solutions to meet the needs and challenges of the organization. (3) Communication and coordination: the enterprise architecture migration group needs to communicate and coordinate closely with the various related teams and stakeholders. They cooperate with business departments, IT departments, data management teams, etc., to jointly formulate migration plans and ensure that parties understand the goals and effects of migration. (4) Performing and monitoring: the enterprise architecture migration group is responsible for implementing the migration plan and monitoring the progress of the overall migration process. They ensure that each task is completed on time, the risk is effectively managed, and change management and problem solving are performed at the same time, so as to ensure smooth proceeding of the migration process. (5) Training and support: an enterprise architecture migration group may need to provide training and support to help employees in an organization adapt to new enterprise architectures. They can provide relevant training materials, guidance and questions to ensure that the organization members can effectively use and maintain the new architecture.

The report subject difference of the bank refers to that report subjects and contents required to be submitted are different among different report objects or institutions. These differences are mainly due to the differences in regulatory body requirements, business type and risk characteristics. The following are some factors that may lead to differences in the delivery subject: (1) regulatory requirements: different regulatory authorities may have different reporting requirements and subject matter of significant concern. For example, a currency policy institution may be more concerned with the bank's capital sufficiency and liquidity, while a financial stabilization institution may be more concerned with the bank's risk management and system importance. (2) Service type: the different types of banks engage in different business areas and therefore the subject matter of their delivery will also vary. Retail banks may be more concerned with the individual deposit and loan situations, while investment banks may be more concerned with the risk of portfolios and trade activities. (3) Risk characteristics: the risk factors faced by banks may vary depending on the regional, market conditions, and economic environment, and thus the subject matter of the report may vary accordingly. For example, certain areas may be more susceptible to natural disasters or political risks, and banks may need to pay special attention to the management of these risks at the time of delivery. (4) Internal management requirements: in addition to regulatory requirements, there may be specific reporting topics and content requirements within the bank. These requirements may be based on the internal risk management framework and strategic goals of the bank, related to capital planning, risk indicators, compliance, etc.

The rule differences of banks refer to differences in operation and management between different banks, which may be due to factors such as legal regulations, internal policies, risk management policies, and organizational cultures. The following are some factors that may lead to differences in bank rules: (1) law and regulation: banks in different countries and regions are subject to different laws and regulations. Financial regulatory authorities in various countries may make different regulations and guidelines to ensure that banks comply with relevant legal regulations in operation. These regulations may relate to capital sufficiency, loan preparation, money back-flushing measures, consumer protection, and the like. (2) Internal policy: each bank will develop a series of internal policies and operational rules based on its own strategic goals and risk management policies. These policies and rules may cover aspects of product design, credit approval process, risk assessment, internal control, and the like. The internal policies of different banks may vary depending on their particular business model and market location. (3) Risk management policy: the preferences and emphasis of risk management may also vary from bank to bank. Some banks may be more focused on credit risk management, while others may be more focused on market risk management or operational risk management. This results in the emphasis and method differences of different banks in formulating risk management policies and procedures. (4) Organizing culture: each bank has unique organization culture and value. These cultural and value views can have an impact on the decisions, management and business practices of the bank. For example, some banks may be more focused on innovation and flexibility, while other banks may be more focused on robustness and conservation.

In order to solve the problem, the embodiment of the application provides a method, a device, equipment and a storage medium for constructing a migration model, which are characterized in that enterprise migration rules, enterprise message processing rules and old architecture core fields are acquired first, and blood-margin analysis technology is utilized to carry out blood-margin analysis on the enterprise migration rules, the enterprise message processing rules and the old architecture core fields to obtain a blood-margin relation set. And then performing migration learning on the blood relationship set by using a migration learning algorithm to obtain a preliminary migration model, and performing migration learning on the first message system by using the migration learning algorithm based on the preliminary migration model to obtain a modified migration model. And then, acquiring the unique theme and the shared theme of the second message system relative to the first message system by using a recommendation system algorithm. And finally, supplementing the transformation migration model by using the second message system, the unique subject and the shared subject to obtain the migration model. According to the application, the blood relationship among the enterprise shelf migration rules, the enterprise shelf message processing rules and the old architecture core fields is traced through the blood relationship analysis technology to obtain a blood relationship set, the blood relationship set and the first message system are subjected to migration learning by using a migration learning algorithm to obtain a modified migration model, and the second message system is analyzed based on a recommendation system algorithm to finally obtain a unique theme and a shared theme of the second message system which can be directly used for supplementing the modified migration model, so that the accuracy and the data quality of enterprise shelf modification can be improved by constructing the migration model, and the communication cost of an enterprise shelf migration group and each project group and the analysis cost of subsequent modification of the project group are reduced.

The following description of the embodiments of the present application will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.

Referring to fig. 1, the method flowchart of a method for constructing a migration model according to an embodiment of the present application, as shown in fig. 1, the method for constructing a migration model may include steps S101 to S105:

s101: and obtaining enterprise frame migration rules, enterprise frame message processing rules and old framework core fields, and performing blood margin analysis on the enterprise frame migration rules, the enterprise frame message processing rules and the old framework core fields by utilizing a blood margin analysis technology to obtain a blood margin relationship set.

In order to construct a migration model, a construction system of the migration model needs to acquire enterprise frame migration rules, enterprise frame message processing rules and old framework core fields, and performs blood margin analysis on the enterprise frame migration rules, the enterprise frame message processing rules and the old framework core fields by using a blood margin analysis technology to obtain a blood margin relationship set.

In some possible implementations, the enterprise migration rules are provided by an enterprise migration group, and the enterprise migration rules are rules for migrating original transaction data to an enterprise under construction. The original transaction data is data which is not processed, and the migration group can process the original transaction data into a migration table, namely an enterprise migration rule according to an enterprise model. For enterprise migration rules, it may be defined according to a specific enterprise architecture and migration goals. The following are examples of some common enterprise migration rules: (1) data mapping and conversion: and determining the mapping relation between the data elements in the old enterprise architecture and the data elements in the new enterprise architecture, and performing corresponding data conversion. This may involve variations in field names, data types, units, standardization, etc. (2) Data integration and reorganization: and integrating and reorganizing the scattered data sources in the old enterprise architecture to adapt to the data integration requirement in the new enterprise architecture. This may require defining a new data model or data table structure and sorting and organizing the data according to the new structure. (3) Interface protocol compatibility: the interface protocol in the new enterprise architecture is ensured to be compatible with the interface protocol in the old enterprise architecture so as to ensure normal communication and data exchange between the systems. This includes coordination and adaptation in terms of protocol format, data transfer, encryption mechanism, etc. (4) Security and rights control: security measures and rights control policies are redesigned and enforced to the security requirements of the new enterprise architecture. This may involve changes in authentication, access control, data encryption, etc. (5) And (3) business process adjustment: and adjusting and redesigning the business processes in the old enterprise architecture according to the business requirements and process optimization targets of the new enterprise architecture. This may require consideration of variations in flow order, data flow, decision routing, etc. (6) System integration and migration sequence: the order and policies of system integration and migration are determined to ensure smooth transitions and functional compatibility of the new enterprise architecture. This may require consideration of factors in terms of dependencies between different systems, interface consistency, etc. (7) Monitoring and feedback mechanisms: and establishing a monitoring and feedback mechanism, detecting problems and risks in the new enterprise architecture in time, and taking corresponding corrective measures. This includes management and optimization in log logging, anomaly alerting, performance monitoring, and the like.

In some possible implementations, a message processing rule refers to a series of rules or steps that process and transform an original message. These rules define how the information in the message is extracted, parsed, converted, and organized to meet specific needs or objectives. Specific messaging rules may include, but are not limited to, the following: (1) message analysis: the original message is parsed in a specific format, such as binary data into structured fields or messages. (2) Extracting fields: extracting the interested field from the analyzed message, and performing extraction operation according to the position, the mark or other characteristics of the field. (3) Field conversion: the extracted fields are converted or mapped to specific data types, units or ranges to meet subsequent processing and analysis requirements. (4) Data cleaning: the extracted fields are subjected to cleaning and preprocessing operations, such as processing missing values, abnormal values, repeated values, noise and the like, so that the quality and the accuracy of data are ensured. (5) Characteristic engineering: the fields are feature engineered, e.g., feature selection, encoding, normalization, dimension reduction, etc., to extract more useful and interpretable features, depending on the specific task or model requirements. (6) Data organization: the processed fields are reorganized into new message formats or data structures to meet the requirements of subsequent processing, storage or transmission. (7) Error handling: for abnormal situations that cannot be resolved or handled, a corresponding error handling policy is defined, such as skip, mark, error reporting, filling default values, etc.

In some possible implementations, the old architecture core field may be, but is not limited to: (1) customer information: including customer name, identification card number, contact address, etc. (2) Account information: including account number, account type (savings account, checking account, etc.), balance, date of account opening, etc. (3) Transaction records: including the amount of money, date, transaction type, etc. for transactions such as deposit, withdrawal, transfer, payment, collection, etc. (4) Loan information: including the loan account number, loan type, loan amount, repayment plan, etc. (5) Payment information: including bank card number, credit card number, payment password, payment status, etc. (6) Safety information: including user login information, access rights, security policies, etc. (7) Statistical data: including transaction amount, deposit total, loan balance, account liveness, etc. The specific core fields may vary from bank to bank business model and system architecture. When performing fabric migration, careful analysis of the data model and business processes in the old fabric is required to determine the exact core fields. Moreover, ensuring data consistency, integrity, and security are also important considerations in the migration of the architecture.

Among these, the blood-margin analysis technique is a technique for understanding the flow and relationship between data. It can help us understand the flow path of data in the system and the correlation with other data by tracking the source of the data, the transfer and conversion process. The following are some common blood-lineage analysis techniques: (1) metadata analysis: attributes, definitions, and usage of data are tracked by analyzing metadata, including data dictionaries, data directories, database structures, and the like. The metadata may provide information about data entities, attributes, relationships, and constraints, thereby helping to determine blood-lineage relationships between data. (2) And (3) log analysis: the flow and processing of data is tracked by analyzing system logs, application logs, database logs, and the like. Operations such as reading, writing, converting and transmitting data are generally recorded in the log, and by analyzing and analyzing the log information, the flow path and the blood-edge relationship of the data can be restored. (3) Data sampling and statistical analysis: the correlation and association between data is obtained by randomly sampling or performing a full statistical analysis of the data. For example, correlation coefficients, frequency numbers, statistical features, etc. between fields may be calculated to reveal blood-vessel relationships between data. (4) ETL tool and data integration platform: modern ETL (extraction, conversion, loading) tools and data integration platforms typically provide blood-margin tracking functionality. They can automatically record the source, conversion, and destination of the data and generate a blood relationship graph or report that helps the analyst understand the flow path of the data. (5) A data management tool: some data governance tools provide functionality for blood-lineage analysis, which can automatically or semi-automatically identify and establish blood-lineage relationships between data. These tools use metadata, logs, etc. information to help the organization manage and maintain data blood-edge relationships.

In some possible implementations, the set of blood-lineage relationships includes: migrating old architecture blood-edge relationships, message old architecture blood-edge relationships, and migrating message blood-edge relationships.

In some possible implementations, migrating old architecture blood-edge relationships refers to the blood-edge logic between enterprise migration rules and old architecture core fields, i.e., [ migration source table, migration source field ] to [ core data table, core field ] blood-edge logic.

In some possible implementations, the old-frame blood-edge relationship of the message refers to the blood-edge logic between the enterprise-frame message processing rule and the old-frame core field, namely, from [ message table, message field ] to [ core data table, core field ].

In some possible implementations, the migration message blood-edge relationship refers to blood-edge logic between the enterprise shelf migration rule and the enterprise shelf message processing rule, that is, from [ message table, message field ] to [ migration source table, migration source field ].

Migrating old architecture blood-edge relationships, message old architecture blood-edge relationships, and migrating message blood-edge relationships can ensure the integrity and accuracy of data.

In some possible implementations, the method further includes:

And uniformly storing the blood edge relations in the blood edge relation set in the source layer database, so that the focus of blood edge analysis can be set on the core table of the source layer database.

The source layer database is used as one layer in the data warehouse and is mainly used for storing the original data extracted from the source system. The core table of the source layer data base contains the most basic and original data information, and has important influence on the whole data flow. The goal of the blood fixing analysis is to track the source and destination of the data, as well as the course of change in the data. By setting the core table of the database of the positioning and source layer, the blood-edge relation of the data can be analyzed more accurately, and the operations comprise data input, output, conversion, processing and the like. The core table is analyzed for blood-fixing reasons, so that people can know the source system, transmission path and use condition of the data by different systems or modules.

S102: and performing migration learning on the blood relationship set by using a migration learning algorithm to obtain a preliminary migration model.

After the blood relationship set is obtained, the migration model building system can utilize a migration learning algorithm to perform migration learning on the blood relationship set to obtain a preliminary migration model.

S103: and performing transfer learning on the first message system by using the transfer learning algorithm based on the preliminary transfer model to obtain a modified transfer model.

After the migration learning algorithm is utilized to perform migration learning on the blood relationship set to obtain a preliminary migration model, the migration model building system can perform migration learning on the first message system based on the preliminary migration model by utilizing the migration learning algorithm to obtain a [ core data table, a core field ] to a [ migration source table, a migration source field ] to a [ message table, a message field ] and a three-section complete transformation migration model.

In some possible implementations, the first messaging system refers to a bank messaging system that performs migration learning first, so as to obtain a modified migration model. A bank messaging system refers to a system within a banking institution for processing and transmitting financial transaction information. The method is responsible for receiving, analyzing, checking and processing various financial transaction messages, extracting data in the messages, and processing and distributing the data according to business rules.

In some possible implementations, the performing the migration learning on the first message system by using the migration learning algorithm based on the preliminary migration model to obtain a modified migration model includes:

A1: and acquiring a message index processing rule, a message index field set and a blood-edge relation between the first message system and the old framework of the first message system.

In order to obtain the transformation migration model, the construction system of the migration model can firstly obtain the message index processing rule, the message index field set and the blood-edge relationship between the first message system and the old framework.

In some possible implementations, the message index processing rule of the first messaging system refers to a rule for processing and analyzing data in a banking transaction message. These rules can be defined and designed according to the specific needs and goals of the bank to extract, transform and calculate key metrics in the message. The following are some common rules for processing the bank message indicators: (1) data extraction rules: these rules define how the required fields or values are extracted from the original message data. For example, specific fields may be extracted using a specified starting position and length, or corresponding data may be extracted by parsing a message in XML or JSON format. (2) Data conversion rules: these rules are used to transform and normalize the extracted data. For example, the monetary amount may be converted from different monetary units to a unified reference monetary unit, or the date-time field may be converted to a standard date-time format. (3) Data calculation rules: these rules define how the required metrics or metrics are calculated from the extracted data. For example, an indicator of total daily transaction amount, average transaction amount, or transaction success rate may be calculated. (4) Data aggregation rules: the rules are used for aggregating the message data according to certain rules. For example, data may be aggregated by dimensions such as customer ID, transaction type, or geographic location to calculate an index in each dimension. (5) Exception handling rules: these rules define how exception or error data is handled. For example, thresholds may be set to detect abnormal transactions, or data verification and error correction may be performed using a reasonable algorithm. (6) Data storage rules: these rules define how the processed index data is stored in the target system or database. For example, data formats, field mappings, and data loading policies may be specified. Equal rules.

In some possible implementations, the set of message indicator fields of the first messaging system may vary according to specific traffic requirements and message formats. The following are some common sets of banking message index fields: transaction Type (Transaction Type): indicating the type of transaction, such as deposit, withdrawal, transfer, payment, etc.; transaction amount (Transaction Amount): a monetary value representing the transaction; transaction Time (Transaction Time): indicating the date and time at which the transaction occurred; transaction status (Transaction Status): representing the status of the transaction, such as success, failure, processing, etc.; account Number (Account Number): a unique identifier representing an account involved in the transaction; customer Number (Customer Number): a unique identifier representing a customer participating in the transaction; transaction location (Transaction Location): indicating the location or channel where the transaction occurred, such as counter, ATM, internet banking, etc.; transaction Fee (Transaction Fee): representing the amount of the commission generated by the transaction; transaction description (Transaction Description): representing descriptive or remark information for the transaction; reference Number (Reference Number): a unique reference number or serial number representing the transaction; counterpart account (Counterparty Account): a unique identifier representing a partner account involved in the transaction; beneficiary Name (Beneficiary Name): representing the beneficiary name or name of the transaction.

In some possible implementations, the first messaging system and old architecture blood-edge relationship refers to a blood-edge relationship of a first messaging system message and an old architecture core field.

A2: and inputting the message index processing rule, the message index field set and the blood relationship between the first message system and the old framework into the preliminary migration model, and performing migration learning by using the migration learning algorithm to obtain a migration model to be tuned.

After the message index processing rule, the message index field set and the blood-edge relation between the first message system and the old framework are obtained, the migration model constructing system can input the message index processing rule, the message index field set and the blood-edge relation between the first message system and the old framework into the preliminary migration model, and the migration learning algorithm is utilized to perform migration learning to obtain the migration model to be tuned.

A3: and carrying out parameter fine adjustment on the migration model to be adjusted to obtain the transformation migration model.

After the migration model to be tuned is obtained, the construction system of the migration model can conduct parameter fine tuning on the migration model to be tuned to obtain the transformation migration model.

The parameter fine tuning is to perform small-amplitude adjustment and optimization on parameters of the model in the migration model training process so as to improve the performance and performance of the model on specific tasks.

The migration model obtained by utilizing the message index processing rule, the message index field set and the blood-margin relation between the first message system and the old framework and through parameter fine adjustment of the first message system can be better suitable for the new field set, the processing rule, the blood-margin relation and the index set after migration, so that the performance and the accuracy of the model are improved.

S104: and acquiring the unique theme and the shared theme of the second message system relative to the first message system by using a recommendation system algorithm.

In some possible implementations, the second message system is a different message system under the same common theme as the first message system, such as the following people-oriented messages: people's bank note number message, people's bank note centralized message, people's bank note interest rate message, etc., the emphasis point of each message is different.

In some possible implementations, the recommender algorithm is a TPLSA-Imp algorithm.

Among them, TPLSA-Imp (Topic-PLSAwith Implicit Feedback) is a migration learning algorithm that combines the ideas of Topic modeling and implicit feedback. The algorithm is used for solving the problem of migration learning in a recommendation system.

S105: and supplementing the transformation migration model by using the second message system, the unique theme and the shared theme to obtain a migration model.

In a possible implementation manner, the supplementing the transformation migration model with the second messaging system, the unique topic and the shared topic to obtain a migration model includes B1-B4:

b1: and acquiring the blood-edge relation between the second message system and the old framework.

In order to supplement the transformation migration model to obtain the migration model, the construction system of the migration model needs to acquire the blood-edge relationship between the second message system and the old framework.

In one possible implementation, the second packet system and old architecture blood-edge relationship refers to the blood-edge relationship of the second packet system and old architecture core field.

B2: and determining an index processing rule of the second message system by utilizing the blood-edge relationship between the second message system and the old framework.

After obtaining the blood-edge relationship between the second message system and the old architecture, the index processing rule of the second message system can be determined by using the blood-edge relationship between the second message system and the old architecture.

The index processing rule is a rule formulated for processing and calculating the original data according to the blood relationship between the second message system and the old framework, the service requirement and the index definition. These rules define how the required data fields are extracted from the system message and the final index result is generated by operations such as arithmetic, aggregation, etc.

B3: and extracting the unique characteristic indexes of the unique subject and the shared characteristic indexes of the shared subject through the index processing rules.

Index processing rules may be applied in index calculations for both unique topics and shared topics. Through index processing rules, we can extract and calculate the unique characteristic index of the unique subject and the shared characteristic index of the shared subject, and further describe and measure their attributes and importance.

B4: and supplementing the transformation migration model by utilizing the unique characteristic indexes and the shared characteristic indexes to obtain a migration model.

The transformation migration model can be supplemented by the second message system, the unique subject and the shared subject, so that the migration model can better capture differences and commonalities among different systems, and the generalization capability and adaptability of the model are improved. The supplementary strategy enables the migration learning to be more flexible and effective, and can better cope with migration tasks and scenes among different systems.

Based on the content of S101-S105, it can be known that the enterprise frame migration rule, the enterprise frame message processing rule and the old architecture core field are obtained first, and the blood margin analysis technology is utilized to perform the blood margin analysis on the enterprise frame migration rule, the enterprise frame message processing rule and the old architecture core field to obtain the blood margin relationship set. And then, performing migration learning on the blood relationship set by using a migration learning algorithm to obtain a preliminary migration model, and performing migration learning on the first message system by using the migration learning algorithm based on the preliminary migration model to obtain a modified migration model. And then, acquiring the unique theme and the shared theme of the second message system relative to the first message system by using a recommendation system algorithm. And finally, supplementing the transformation migration model by using the second message system, the unique subject and the shared subject to obtain the migration model. According to the application, the blood relationship among the enterprise shelf migration rules, the enterprise shelf message processing rules and the old architecture core fields is traced through the blood relationship analysis technology to obtain a blood relationship set, the blood relationship set and the first message system are subjected to migration learning by using a migration learning algorithm to obtain a modified migration model, and the second message system is analyzed based on a recommendation system algorithm to finally obtain a unique theme and a shared theme of the second message system which can be directly used for supplementing the modified migration model, so that the accuracy and the data quality of enterprise shelf modification can be improved by constructing the migration model, and the communication cost of an enterprise shelf migration group and each project group and the analysis cost of subsequent modification of the project group are reduced.

Referring to fig. 2, fig. 2 is a schematic structural diagram of a migration model building apparatus according to an embodiment of the present application. As shown in fig. 2, the migration model building apparatus includes:

the obtaining and analyzing unit 201 is configured to obtain an enterprise frame migration rule, an enterprise frame message processing rule, and an old architecture core field, and perform blood edge analysis on the enterprise frame migration rule, the enterprise frame message processing rule, and the old architecture core field by using a blood edge analysis technology to obtain a blood edge relationship set.

In some possible implementations, the apparatus further includes:

and the storage unit is used for storing the blood relationship in the blood relationship set in the source layer database.

The first learning unit 202 is configured to perform transfer learning on the blood relationship set by using a transfer learning algorithm to obtain a preliminary transfer model.

The second learning unit 203 is configured to perform transfer learning on the first message system by using the transfer learning algorithm based on the preliminary transfer model to obtain a modified transfer model.

In some possible implementations, the apparatus further includes:

the second acquisition unit is used for acquiring the message index processing rule, the message index field set and the blood-edge relationship between the first message system and the old framework of the first message system.

The input unit is used for inputting the message index processing rule, the message index field set and the blood relationship between the first message system and the old framework into the preliminary migration model, and performing migration learning by utilizing the migration learning algorithm to obtain a migration model to be tuned.

The first obtaining unit 204 is configured to obtain, using a recommender algorithm, a unique topic and a shared topic of the second messaging system relative to the first messaging system.

Among them, TPLSA-Imp (Topic-PLSA with Implicit Feedback) is a migration learning algorithm that combines the ideas of Topic modeling and implicit feedback. The algorithm is used for solving the problem of migration learning in a recommendation system.

The first supplementing unit 205 is configured to supplement the transformation migration model with the second message system, the unique theme, and the shared theme to obtain a migration model.

In one possible implementation, the apparatus further includes:

and the third acquisition unit is used for acquiring the blood-edge relation between the second message system and the old framework.

And the determining unit is used for determining the index processing rule of the second message system by utilizing the blood-edge relation between the second message system and the old framework.

And the extraction unit is used for extracting the unique characteristic indexes of the unique topics and the shared characteristic indexes of the shared topics through the index processing rules.

And the second supplementing unit is used for supplementing the transformation migration model by utilizing the unique characteristic indexes and the shared characteristic indexes to obtain a migration model.

The embodiment of the application provides a device for constructing a migration model, which comprises the steps of firstly acquiring enterprise frame migration rules, enterprise frame message processing rules and old framework core fields by using an acquisition and analysis unit 201, and carrying out blood margin analysis on the enterprise frame migration rules, the enterprise frame message processing rules and the old framework core fields by using a blood margin analysis technology to obtain a blood margin relation set. The first learning unit 202 performs transfer learning on the blood relationship set by using a transfer learning algorithm to obtain a preliminary transfer model, so that the second learning unit 203 may perform transfer learning on the first message system by using the transfer learning algorithm based on the preliminary transfer model to obtain a modified transfer model. The first obtaining unit 204 obtains the unique theme and the shared theme of the second message system relative to the first message system by using a recommendation system algorithm, so that the first supplementing unit 205 can supplement the transformation migration model by using the second message system, the unique theme and the shared theme to obtain the migration model. According to the application, the blood relationship among the enterprise shelf migration rules, the enterprise shelf message processing rules and the old architecture core fields is traced through the blood relationship analysis technology to obtain a blood relationship set, the blood relationship set and the first message system are subjected to migration learning by using a migration learning algorithm to obtain a modified migration model, and the second message system is analyzed based on a recommendation system algorithm to finally obtain a unique theme and a shared theme of the second message system which can be directly used for supplementing the modified migration model, so that the accuracy and the data quality of enterprise shelf modification can be improved by constructing the migration model, and the communication cost of an enterprise shelf migration group and each project group and the analysis cost of subsequent modification of the project group are reduced.

The method, the device, the equipment and the storage medium for constructing the migration model provided by the application are described in detail. In the description, each embodiment is described in a progressive manner, and each embodiment is mainly described by the differences from other embodiments, so that the same similar parts among the embodiments are mutually referred. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section. It should be noted that it will be apparent to those skilled in the art that various modifications and adaptations of the application can be made without departing from the principles of the application and these modifications and adaptations are intended to be within the scope of the application as defined in the following claims.

It should be noted that, the method, the device, the equipment and the storage medium for constructing the migration model provided by the present application may be used in the big data field or the financial field, and the above is only an example, and the method, the device, the equipment and the application field of the storage medium for constructing the migration model provided by the present application are not limited.

It should be understood that in the present application, "at least one (item)" means one or more, and "a plurality" means two or more. "and/or" for describing the association relationship of the association object, the representation may have three relationships, for example, "a and/or B" may represent: only a, only B and both a and B are present, wherein a, B may be singular or plural. The character "/" generally indicates that the context-dependent object is an "or" relationship. "at least one of" or the like means any combination of these items, including any combination of single item(s) or plural items(s). For example, at least one (one) of a, b or c may represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", wherein a, b, c may be single or plural.

It is further noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. The software modules may be disposed in Random Access Memory (RAM), memory, read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

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.

Claims

1. The method for constructing the migration model is characterized by comprising the following steps:

2. The method of claim 1, wherein performing the migration learning on the first message system by using the migration learning algorithm based on the preliminary migration model to obtain the modified migration model comprises:

3. The method of claim 1, wherein the set of blood-lineage relationships includes: migrating old architecture blood-edge relationships, message old architecture blood-edge relationships, and migrating message blood-edge relationships.

4. A method according to claim 1 or 3, characterized in that the method further comprises:

5. The method of claim 1, wherein the recommender system algorithm is a TPLSA-Imp algorithm.

6. The method of claim 1, wherein supplementing the remodelled migration model with the second messaging system, the unique topic, and the shared topic results in a migration model, comprising:

7. A migration model building apparatus, wherein the apparatus comprises:

8. The apparatus of claim 7, wherein the apparatus further comprises:

9. A migration model construction apparatus, comprising: memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the method of constructing a migration model according to any one of claims 1-6 when the computer program is executed.

10. A computer readable storage medium, characterized in that the computer readable storage medium has stored therein instructions, which when run on a terminal device, cause the terminal device to perform the method of constructing a migration model according to any one of claims 1-6.