A standardized management method for bank data

Abstract

The invention discloses a standardized management method for bank data, belonging to the field of bank data management. A standardized management method for bank data includes the following steps: S10. Extracting a root word library; S20. Combing a business word library; S30. Standardizing the Chinese standard naming of fields; S40. Establishing a mapping relationship between the original Chinese name of the field and the standardized naming; S50. Forming an enterprise-level standardized data dictionary; It can promote enterprise-level data homology management, ensure the consistency and sharing of data across systems and platforms, and make data easy to identify, reference and share by establishing a data standard management system, so as to improve the accuracy and availability of data.

Description

Technical Field

[0001] The present invention belongs to the field of bank data management, and more specifically, relates to a standardized management method for bank data. Background Art

[0002] The banking industry faces a series of challenges in its digital transformation and in the process of realizing the value of data. Different information system construction backgrounds, management requirements, and data demands lead to ambiguous data semantics, and irregular phenomena such as synonyms but different names, and the same name but different meanings. This not only affects the accuracy and availability of the data, but also brings difficulties to data standard management.

[0003] Realizing unified management of data standards and promoting strict implementation of data standards are difficult issues that the banking industry urgently needs to solve. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a standardized management method for bank data, which can promote enterprise-level data homogeneity management, ensure the consistency and sharing of data across systems and platforms, and make data easy to identify, reference and share by establishing a data standard management system, thereby improving the accuracy and availability of data.

[0005] A standardized management method for bank data of the present invention comprises the following steps:

[0006] S10. Extract root word database;

[0007] S20. Combing the business term database;

[0008] S30. Standardize the Chinese standard naming of fields;

[0009] S40. Establish a mapping relationship between the original field Chinese name and the standardized name;

[0010] S50. Form an enterprise-level standardized data dictionary.

[0011] As a further improvement of the present invention, the specific steps of refining the root word library include:

[0012] S11. Input the Chinese names of the fields and related information from the business table and the derived table in the metadata management information system; split the Chinese names of the metadata into words according to the dimension of natural language;

[0013] S12. The phrases formed by the word splitting are judged according to their attributes and divided into root words, modifier words, and general words;

[0014] S13. Check whether the split words meet the root word requirements, and divide the review results into "in use", "not applicable", and "pending review". For the "pending review" words, select the words that meet the business scenarios, meet the root word feature requirements, and meet the semantic expression specifications as roots to form a root word library.

[0015] As a further improvement of the present invention, the modifiers and roots obtained by word decomposition are intelligently combined in pairs and effectively paired according to established semantic rules and logical relationships; whether the phrases formed by the combination follow the standard format of "modifier + root" is reviewed to form an initial business term library.

[0016] As a further improvement of the present invention, the attributes of the entries formed by word grouping are divided into three states: "in use", "need manual review", and "not applicable"; if it does not end with a root word and does not have actual business meaning, it is judged to be "not applicable"; if it meets the format requirements and has actual business meaning, it is judged to be "in use"; if it meets the format requirements but the business meaning is unclear, it is judged to be "need manual review"; for the phrases marked as "need manual review" by the system, the data governance personnel will conduct further review to ensure that the naming of the phrases is both standardized and in line with the business demand logic.

[0017] As a further improvement of the present invention, the entry library supports manual addition, modification, query, and deletion of entries; if the existing entry cannot match the Chinese name of the newly added field, the business entry is expanded; if the business is disabled, it is necessary to evaluate whether to disable the relevant business entries and roots.

[0018] As a further improvement of the present invention, the Chinese names of fields in different business scenarios are standardized through the key index of business terms; based on the root library and the business term library, the Chinese names of fields in the information system are reviewed, and the review results are divided into "standard" and "non-standard"; the data governance team reviews the fields with "non-standard" review results to form standardized naming of Chinese fields.

[0019] As a further improvement of the present invention, the Chinese names of the fields of each business theme are counted, and the Chinese names of the fields of the business table and the derived table are standardized in combination with the root library and the business term library, and a mapping relationship between the original field Chinese names and the standardized names is established.

[0020] As a further improvement of the present invention, through a comprehensive inventory of various business systems, standardized Chinese field management is implemented to form an enterprise-level standardized data dictionary; based on the standardized naming of the fields, the unique keyword attributes in the business terms are utilized to aggregate and form a combination of similar business data, clarify the authoritative owner and the business owner, and establish a data relationship; for the fields of the authoritative owner, a business data standard is formed, and the business owner shall refer to and implement the standard requirements of the authoritative owner.

[0021] Compared with the prior art, the beneficial effects of the present invention are as follows: by defining data standards, clarifying the management attributes, business attributes, and technical attributes of the standards, the Chinese names of the database table fields are standardized, and the business significance of the data can be fully and accurately expressed, thereby promoting the improvement of data management efficiency;

[0022] By reviewing and combing through data from previous years and standardizing field naming, we have promoted the iterative upgrade of standards from department level to enterprise level, promoted enterprise-level data homogeneity management, and ensured the consistency and sharing of data across systems and platforms. By establishing a data standard management system, we make data easy to identify, reference, and share, enabling business departments to improve data application efficiency and development departments to improve system development efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a schematic diagram of the standardized name splitting and attribute tagging operation flow of the present invention;

[0024] Figure 2 It is a schematic diagram of the word formation process of the present invention;

[0025] Figure 3 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0026] Specific embodiment 1: Please refer to Figure 1-Figure 3 A standardized management method for bank data includes the following steps:

[0027] S10. Extract root word database;

[0028] A root word refers to a word that can express a data type, and a modifier is a word that modifies the root word in order to express the meaning of the data more clearly. For example, "currency code" is composed of the modifier "currency" and the root word "code", and "name of the school where you graduated" is composed of the modifiers "graduation", "school" and the root word "name".

[0029] In this step, the Chinese names of the fields of the stock information system need to be split into three categories: modifiers, roots, and general. General refers to words that can be used as both modifiers and roots.

[0030] The specific splitting method is as follows:

[0031] S11. Input the Chinese names of the fields and related information of the business table and derived table from the metadata management information system (business table refers to the table used to record the business process, and derived table refers to the table processed based on the business table). Split the Chinese names of metadata into words according to the dimension of natural language through the jieba Chinese word segmentation library. For example, split "currency code" into "currency" and "code"; split "yesterday's balance direction code" into "yesterday", "balance", "direction", and "code".

[0032] S12. The phrases formed by the word splitting are judged according to their attributes and divided into root words, modifiers, and general words. A root word can have multiple modifiers at the same time, and usually 1-3 modifiers are used. Root words refer to words that can express data types. General words refer to words that can be used as both root words and modifiers. For example, after the "currency code" is split, "currency" is a modifier, and "code" is general; after the "yesterday's balance direction code" is split, "yesterday" is a modifier, "balance" is general, "direction" is a modifier, and "code" is general.

[0033] S13. Check whether the split words meet the root requirements, and divide the audit results into "in use", "not applicable", and "pending review"; "in use" means that the word meets the root definition, "not applicable" means that the word does not meet the root definition, and "pending review" means that the word needs to be manually judged whether it is a root. For the "pending review" words, data governance personnel select words that meet the business scenarios, meet the root feature requirements, and meet the semantic expression specifications as roots to form a root library. Referring to the classification of data types in the People's Bank of China's "JR / T 0105-2014 Bank Data Standard Definition Specification", the roots are divided into 10 categories: ratio, code, coding, sign, amount, date, time, date and time, value, and text. The root library supports iterative updates caused by operations such as adding a new system and modifying standardized names, and supports manual addition, modification, query, and deletion of data elements.

[0034] S20. Combing the business term database;

[0035] Business terms are basic descriptions of business elements, defining unique keywords for a certain type of business data. The combination consists of "modifier + root word", ensuring the accuracy and consistency of data naming.

[0036] Business terms cover data elements of business subjects such as deposits, loans, and funds, such as deposit balances, loan amounts, and fund accounts. Business-specific names and system-specific terms are also included in the term library management but no format verification is performed, such as "registered capital" and "self-increment number".

[0037] Word grouping refers to the intelligent combination of modifiers and word roots obtained by word decomposition, and effective pairing according to established semantic rules and logical relationships. Then, the combined phrases are reviewed to see whether they follow the standard format of "modifier + word root" to form an initial business term library.

[0038] The specific steps of word formation are:

[0039] For example, the term "yesterday's balance direction code" can be divided into four words: "yesterday", "balance", "direction" and "code". The above four words can be combined into words such as "yesterday's balance", "balance direction" and "direction code".

[0040] The term "currency code" can be divided into the words "currency" and "code". The above two words can be combined into the term "currency code".

[0041] The properties of the entries formed by word combinations are divided into three states: "in use", "requires manual review", and "not applicable". If the entry does not end with a root word and has no actual business meaning, it is judged as "not applicable"; if it meets the format requirements and has actual business meaning, it is judged as "in use"; if it meets the format requirements but the business meaning is unclear, it is judged as "requires manual review". For the phrases marked as "requires manual review" by the system, data governance personnel will conduct further reviews to ensure that the naming of the phrases is both standardized and in line with the logic of business needs.

[0042] The tool judgment logic is shown in Table 1. The entry library supports manual addition, modification, query, and deletion of entries. If the existing entry cannot match the Chinese name of the newly added field, the business entry is expanded; if the business is disabled, it is necessary to evaluate whether to disable the relevant business entries and roots.

[0043] Table 1 Tool judgment logic table

[0044]

[0045] S30. Standardize the Chinese standard naming of fields;

[0046] Standardized naming means naming fields in a prescribed format in the basic form of "modifier (can be multiple) + business terms". Business terms can be used separately as standardized names for separate Chinese fields. Through the key index of business terms, the Chinese names of fields in different business scenarios are standardized. For example, for the business term "customer name", in different business scenarios, the standardized Chinese names of fields are named "customer name", "legal representative customer name", "beneficiary customer name", "guarantor customer name", and "corporate customer name". Based on the root word library and business term library, the Chinese names of fields in the information system are reviewed, and the review results are divided into "standard" and "non-standard". The data governance team reviews the fields with "non-standard" review results to form standardized naming for Chinese fields.

[0047] S40. Establish a mapping relationship between the original field Chinese name and the standardized name;

[0048] Take inventory of the Chinese names of the fields of each business theme, standardize the Chinese names of the fields in the business table and derived tables in combination with the root word library and business term library, and establish a mapping relationship between the original field Chinese names and the standardized names.

[0049] Standardized naming enhances the readability and usability of fields, improves the consistency and accuracy of data applications between systems and platforms, reduces communication costs, and improves communication efficiency; at the same time, the mapping relationship between the original field Chinese name and the standardized name provides a clear reference for data migration, integration, and analysis; through the data mapping relationship, the implementation of enterprise-level standards is promoted, avoiding the transformation of related original system tables. The mapping relationship implements dynamic supervision and management according to the changes in the system or table fields.

[0050] S50. Form an enterprise-level standardized data dictionary;

[0051] Through a comprehensive inventory of each business system, we implement standardized Chinese field management, strengthen cross-business, cross-system, and cross-platform data integration, and form an enterprise-level standardized data dictionary. Based on the standardized naming of fields, we use the unique keyword attributes in business terms to aggregate and form a combination of similar business data, clarify the authoritative owner and business owner (the meaning of the authoritative owner and business owner has been explained in the existing patent with patent number CN117114843B, so I won’t repeat it here), and establish data relationships. For the fields of the authoritative owner, further clarify the management attributes, business attributes, and technical attributes of the data to form business data standards. Business owners should refer to and implement the standard requirements of the authoritative owner. By clarifying the interactive relationship between data, we can ensure the accurate flow of data across businesses and systems, enhance the consistency and traceability of data, and promote the continuous improvement of data quality.

[0052] The above data management methods can be applied in the following scenarios:

[0053] 1. Application in major system development (or system transformation).

[0054] 1. It can implement standard recommendations. It recommends available root words, business terms, and standardized names, providing the project team with a set of clear references for field Chinese names, which helps the project team quickly select and apply standardized field names during the development process.

[0055] Second, it can achieve standard audit. After the project team forms the library and table definitions, the data governance team will check the Chinese names of the fields in the library and table structures to ensure that the Chinese names of the fields follow the established naming rules and business logic, and to ensure the standardization and consistency of the data.

[0056] Third, it can achieve standard dynamic management. For newly added business elements, after discussion and confirmation by the data governance team, business department, and technology department, the root word library and business term library are expanded and updated to ensure that the root word library and business term library adapt to new business needs.

[0057] 2. Application in data systems.

[0058] Data standards play a key role in serving data systems. For data development services, developers can quickly select the data items they need to use for report definition, self-service analysis, and data modeling by using the standardized naming of data dictionaries, thereby improving development efficiency and ensuring the consistency and accuracy of data during the development process. For business personnel, data standards make data easy to identify and reference, greatly reducing the communication costs caused by differences in data understanding. Through standardized naming and clear data definitions, business personnel can quickly and easily find the required data, conduct effective data analysis and decision support, and improve data application efficiency.

[0059] 3. Application in data quality monitoring.

[0060] Through the uniqueness of business terms, the relationship between the authoritative owner and the business owner is clarified, effectively solving the problem of cross-system data consistency. For authoritative owner data, the data standard clarifies the business attributes, technical attributes, and management attributes, so that data quality problems can be identified, warned, and fed back early, promoting the rectification of data quality problems and the optimization of information systems, strengthening data source control, and improving data quality.

[0061] Specific embodiment 2: Based on the specific embodiment 1, the specific steps of forming words also include:

[0062] Retrieve the root word in the split word, lock the root word, and trace the category corresponding to the root word according to Table 2; trace the metadata of the source of the root word again, lock the modifiers that match the root word in all metadata, and the so-called match means that it appears in the same phrase with the root word; combine the locked modifier word with the root word, and the combination method is at least one modifier word + one root word, and the root word is the end of the phrase; identify the new phrase combined in the above method as a "used" phrase, update it into the root word library, and match the corresponding category.

[0063] Retrieve the "common" word from the split words, and trace the category corresponding to the "common" word according to Table 2; trace the metadata of the source of the "common" word again, and lock the modifiers that match the "common" word in all metadata, where the so-called match means that they appear in the same phrase as the "common" word; group the locked modifiers with the "common" word in a way of at least one modifier + one "common" word, with the "common" word as the end of the phrase; identify the new phrase combined in the above way as a "used" phrase, update it into the root word library, and match the corresponding category.

[0064] It is prohibited to form words ending with modifiers, and all new words formed in the above way will be eliminated.

[0065] The "common" words and word roots retrieved above are combined and matched to form new phrases, and the new phrases formed in the above manner are packaged and sent to manual review; wherein, the ending word is used as the basis for category identification, and the new phrases are packaged separately by category.

[0066] Table 2 Root word corresponding category tracing table

[0067]

[0068] Specific embodiment three: On the basis of specific embodiment one or two, when standardizing the Chinese standard naming of the field, the last word of the phrase in the entry library is set as the A-position word, and the word before the last word is set as the B-position word. The target phrase is locked, and the combination of "B-position word + A-position word" is searched in the metadata. If the probability of its occurrence reaches the set ratio, it is identified as a standard business entry and can be treated as a separate phrase. In addition, a new position word can be set in front of the B-position word to supplement the identification of the standard business entry.

Claims

1. A standardized management method for bank data, characterized by: The following steps are involved: S10. Extract the root word library, and the specific steps include: S11. Input the Chinese names of the fields and related information of the business table and the derived table from the metadata management information system; split the metadata Chinese names into words according to the dimension of natural language; S12. Judge the phrases formed by the split words according to the attributes and divide them into roots, modifiers, and general words; S13. Review whether the split words meet the root word requirements, and divide the review results into "in use", "not applicable", and "pending review"; for the words "pending review", select the words that meet the business scenarios, meet the root word feature requirements, and meet the semantic expression specifications as roots to form a root word library; S20. Sort out the business term library; business terms are basic descriptions of business elements, defining the unique keywords of a certain type of business data, and the combination form is composed of "modifiers + root words"; word combination refers to the intelligent combination of modifiers and root words obtained by word decomposition, and effective pairing according to the established semantic rules and logical relationships, and then review whether the combined phrases follow the "modifiers + root word" standard format to form an initial business term library; S30. Standardize the Chinese naming of fields; standardized naming means naming fields in the basic form of "modifier + business terms" through a prescribed format, and business terms can be used separately as standardized naming of separate Chinese fields; standardize the Chinese names of fields in different business scenarios through the key index of business terms; review the Chinese names of fields in the information system based on the root word library and business term library, and the review results are divided into "standard" and "non-standard"; the data governance team reviews the fields with the review results of "non-standard" to form standardized naming of Chinese fields; S40. Inventory the Chinese names of the fields of each business theme, combine the root word library and the business term library, standardize the Chinese names of the fields in the business table and the derived table, and establish a mapping relationship between the original field Chinese name and the standardized name; S50. Form an enterprise-level standardized data dictionary; through a comprehensive inventory of various business systems, implement standardized Chinese field management and form an enterprise-level standardized data dictionary; based on the standardized naming of fields, use the unique keyword attributes in business terms to aggregate and form a combination of similar business data, clarify the authoritative owner and business owner, and establish data relationships; for the fields of the authoritative owner, form business data standards, and the business owner shall refer to and implement the standard requirements of the authoritative owner.

2. A standardized management method for bank data according to claim 1, characterized in that: The properties of the entries formed by word groups are divided into three states: "in use", "need manual review", and "not applicable"; if they do not end with a root word and do not have actual business meaning, they are judged to be "not applicable"; if they meet the format requirements and have actual business meaning, they are judged to be "in use"; if they meet the format requirements but the business meaning is unclear, they are judged to be "need manual review"; for phrases marked as "need manual review" by the system, data governance personnel will conduct further reviews to ensure that the naming of the phrases is both standardized and in line with business demand logic.

3. A standardized management method for bank data according to claim 1, characterized in that: The entry database supports manual addition, modification, query, and deletion of entries. If the existing entry cannot match the Chinese name of the newly added field, the business entry will be expanded. If the business is disabled, it is necessary to evaluate whether to disable the relevant business entries and roots.

Images