A core field identification method and system

Abstract

The embodiment of the application provides a core field identification method and system, which can utilize a pre-trained identification model to identify each to-be-identified field in a to-be-identified table. In the case that the identification result indicates that the to-be-identified field is a core field, and the field names of the core fields corresponding to the to-be-identified field are multiple, the field name of a target core field with the highest matching degree with the to-be-identified field is determined from the multiple field names of the core fields based on the information of the to-be-identified table. The whole method can not only quickly identify the to-be-identified field with high efficiency, but also can uniformly name each to-be-identified field because the field names of the core fields are pre-annotated, thereby improving the data use efficiency and laying a foundation for cross-department data linkage analysis.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a core field identification method and system. Background Technology

[0002] As steel manufacturing enterprises continue to develop, various departments within these enterprises will establish corresponding information systems based on their business needs, such as office automation (OA) systems, attendance management systems, material management systems, and vehicle management systems. Each information system has a corresponding database and multiple tables that use these databases. Each table is used to summarize or display different data, and each table has multiple fields depending on the different data being summarized.

[0003] However, due to the lack of a unified field naming standard, fields with the same meaning may have different names in different tables across different departments' information systems. For example, "Factory," "Factory Code," "Production Plant," "Process Code," "compy_cd," "cmp," "plt," "factory," and "fac_cd" all represent the "Factory" field. Inconsistent field names reduce data utilization efficiency and hinder cross-departmental data collaboration and analysis. Furthermore, because steel manufacturing enterprises experience frequent process evolution, equipment iteration, and raw material replacement, various information systems retain a large number of past fields. These past fields are irrelevant to current production activities and are therefore rarely used in the current production process. However, for data analysis, it is necessary to identify all fields with different names but representing the same meaning from various information systems.

[0004] Currently, field identification is mainly achieved by identifying and filtering all fields stored in various databases of various information systems one by one. The inconsistent field names and the existence of a large number of irrelevant fields not only make the identification workload large, but also make the identification efficiency very low. Summary of the Invention

[0005] This application provides a core field recognition method and system, which can be used to solve the technical problems of existing field recognition methods having a large workload and low recognition efficiency.

[0006] In a first aspect, embodiments of this application provide a core field identification method, including:

[0007] Obtain the table to be identified, which includes multiple fields to be identified;

[0008] Multiple fields to be identified are input into the recognition model for core field recognition to obtain the recognition results of each field to be identified. The recognition model is trained and validated using the labeled core fields.

[0009] If the recognition result of the field to be identified indicates that the field to be identified is a core field, then obtain the labeled field name of the core field corresponding to the field to be identified;

[0010] If there are multiple labeled field names for the core field corresponding to the field to be identified, then based on the information of the table to be identified, the labeled field name of the target core field with the highest matching degree with the field to be identified is determined from the labeled field names of the multiple core fields corresponding to the field to be identified.

[0011] Rename the field name of the field to be identified to the label field name of the target core field.

[0012] In conjunction with the first aspect, one possible implementation of the first aspect also includes:

[0013] If the label field name of the core field corresponding to the field to be identified is the same, then the field name of the field to be identified is renamed to the label field name of the core field.

[0014] In conjunction with the first aspect, in one possible implementation of the first aspect, the step of inputting multiple fields to be identified into a recognition model for core field recognition to obtain recognition results for each of the fields to be identified includes:

[0015] Extract the features of each of the fields to be identified;

[0016] The features of the field to be identified are compared with the features of each core field obtained in advance to determine the similarity between the field to be identified and each core field.

[0017] If there is a core field whose similarity to the field to be identified is greater than a preset threshold, then the field to be identified is determined to be a core field.

[0018] In conjunction with the first aspect, in one possible implementation of the first aspect, the extraction of features from each of the fields to be identified includes:

[0019] By scanning the field names and field descriptions of each field to be identified using natural language processing technology, the features of each field to be identified are obtained.

[0020] In conjunction with the first aspect, in one possible implementation of the first aspect, the features of the field to be identified are compared with the features of each pre-acquired core field to determine the similarity between the field to be identified and each core field, including:

[0021] The k-means algorithm is used to cluster the features of the field to be identified and the features of each core field to obtain the clustering results of the field to be identified.

[0022] If the clustering result of the field to be identified contains a core field, then the similarity between the field to be identified and the core field is obtained.

[0023] In conjunction with the first aspect, in one possible implementation of the first aspect, the information of the table to be identified includes at least one of the following: the field value of each field to be identified, the annotation of each field to be identified, the table name of the table to be identified, and the annotation of the table to be identified.

[0024] In conjunction with the first aspect, in one possible implementation of the first aspect, the core field is a field strongly related to current production, and the labeled core field is obtained in advance through the following method:

[0025] Obtain the frequency of occurrence of each field in the information system;

[0026] Fields that appear more frequently than a preset frequency are identified as candidate fields;

[0027] Identify the core fields from among the candidate fields;

[0028] The field names of the core fields are labeled.

[0029] In conjunction with the first aspect, one possible implementation of the first aspect also includes:

[0030] Based on the recognition results of each of the fields to be identified, identify the incorrect fields that do not conform to the core fields and the missing fields that did not identify the core fields from each of the fields to be identified;

[0031] Based on the identified error fields, the missed identification fields, and the pre-acquired core fields, the identification model is retrained and the parameters of the identification model are updated.

[0032] Secondly, embodiments of this application provide a core field recognition system, including:

[0033] A field to be identified acquisition unit is used to acquire a table to be identified, wherein the table to be identified includes multiple fields to be identified;

[0034] The core field recognition unit is used to input multiple fields to be recognized into the recognition model to perform core field recognition and obtain the recognition results of each field to be recognized. The recognition model is trained and verified using the labeled core fields.

[0035] The core field acquisition unit is used to acquire the labeled field name of the core field corresponding to the field to be identified if the recognition result of the field to be identified indicates that the field to be identified is a core field.

[0036] The target core field determination unit is used to determine the label field name of the target core field that has the highest matching degree with the target core field from the label field names of the multiple core fields corresponding to the target field, based on the information of the table to be identified, if there are multiple field names of the core field corresponding to the target field.

[0037] The renaming unit is used to rename the field name of the field to be identified to the label field name of the target core field.

[0038] In conjunction with the second aspect, in one possible implementation of the second aspect, the renaming unit is also used for:

[0039] If the label field name of the core field corresponding to the field to be identified is the same, then the field name of the field to be identified is renamed to the label field name of the core field.

[0040] In conjunction with the second aspect, in one possible implementation of the second aspect, the core field identification unit is specifically used for:

[0041] Extract the features of each of the fields to be identified;

[0042] The features of the field to be identified are compared with the features of each core field obtained in advance to determine the similarity between the field to be identified and each core field.

[0043] If there is a core field whose similarity to the field to be identified is greater than a preset threshold, then the field to be identified is determined to be a core field.

[0044] In conjunction with the second aspect, in one possible implementation of the second aspect, the core field identification unit is specifically used for:

[0045] By scanning the field names and field descriptions of each field to be identified using natural language processing technology, the features of each field to be identified are obtained.

[0046] In conjunction with the second aspect, in one possible implementation of the second aspect, the core field identification unit is specifically used for:

[0047] The k-means algorithm is used to cluster the features of the field to be identified and the features of each core field to obtain the clustering results of the field to be identified.

[0048] If the clustering result of the field to be identified contains a core field, then the similarity between the field to be identified and the core field is obtained.

[0049] In conjunction with the second aspect, in one possible implementation of the second aspect, the information of the table to be identified includes at least one of the following: the field value of each field to be identified, the annotation of each field to be identified, the table name of the table to be identified, and the annotation of the table to be identified.

[0050] In conjunction with the second aspect, in one possible implementation of the second aspect, the core field is a field strongly related to the current production, and the labeled core field is obtained in advance through the following method:

[0051] Obtain the frequency of occurrence of each field in the information system;

[0052] Fields that appear more frequently than a preset frequency are identified as candidate fields;

[0053] Identify the core fields from among the candidate fields;

[0054] The field names of the core fields are labeled.

[0055] In conjunction with the second aspect, one possible implementation of the second aspect further includes a model update unit, which is used for:

[0056] Based on the recognition results of each of the fields to be identified, identify the incorrect fields that do not conform to the core fields and the missing fields that did not identify the core fields from each of the fields to be identified;

[0057] Based on the identified error fields, the missed identification fields, and the pre-acquired core fields, the identification model is retrained and the parameters of the identification model are updated.

[0058] This application provides a core field identification method and system. The core field identification method utilizes a pre-trained identification model to identify each field in a table to be identified. If the identification result indicates that a field to be identified is a core field, and there are multiple core field names corresponding to the field to be identified, then the target core field name with the highest matching degree is determined from the multiple core field names based on the information of the table to be identified. The entire method uses an identification model to determine whether each field to be identified is a core field, and uses the field name of the corresponding core field to rename the field to be identified. This not only enables rapid and efficient identification of the field to be identified, but also allows for standardized and unified naming of each field to be identified since the core field names have been pre-labeled, thereby improving data utilization efficiency and laying the foundation for cross-departmental data linkage analysis. Attached Figure Description

[0059] Figure 1This is a schematic diagram illustrating the workflow of a core field identification method provided in some embodiments of this application;

[0060] Figure 2 This is a schematic diagram of the workflow of a core field identification method provided in some other embodiments of this application;

[0061] Figure 3 This is a schematic diagram of the structure of a core field recognition system provided in an embodiment of this application. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0063] The following section first introduces the possible scenarios to which the embodiments of this application are applicable.

[0064] Within steel manufacturing enterprises, each department establishes corresponding information systems based on its business needs, such as OA systems, attendance management systems, material management systems, and vehicle management systems. Each information system contains a corresponding database and multiple tables that use these databases. For example, the vehicle management system includes tables for basic population information, basic vehicle information, and recharge records. Each table has multiple fields; for instance, the table for basic population information includes fields such as name, age, ID number, ID type, home address, and contact number.

[0065] Due to the lack of a unified field naming standard, fields with the same meaning in different departments' information systems may have different names in different tables. For example, see Table 1 below.

[0066] Table 1: Comparison of field names for the same field in different tables

[0067]

[0068] Inconsistent and disorganized field names can cause significant confusion for business users, reduce data utilization efficiency, and hinder cross-departmental data collaboration and analysis. Due to the existence of numerous past records, each information system may contain thousands of tables, each with hundreds or even thousands of fields. Traditional field identification primarily involves identifying and filtering each field individually, which is not only labor-intensive but also extremely inefficient.

[0069] To address the technical problems of traditional identification methods, this application discloses a core field identification method through the following embodiments. The core field identification method provided in this application utilizes a pre-trained identification model to identify each field to be identified. If a field to be identified is identified as a core field, the field name of the field to be identified is renamed using the field name of the corresponding core field. This not only improves identification efficiency but also allows for standardized naming of the fields to be identified, which is beneficial for improving data utilization efficiency. See also Figure 1 The flowchart shown illustrates a core field identification method provided in some embodiments of this application, which specifically includes the following steps:

[0070] 101: Obtain the table to be recognized.

[0071] The table to be identified includes multiple fields to be identified, such as the row labels and column labels in the table.

[0072] In some embodiments, there can be multiple tables to be identified. For example, if a user selects a database within an information system, all tables associated with that database are identified as tables to be identified, and the total number of fields to be identified in that database is the sum of the fields in all tables associated with that database. As another example, if a user selects an information system, all tables associated with each database within that information system are identified as tables to be identified, and the total number of fields to be identified in that information system is the sum of the fields in all tables associated with each database.

[0073] In other embodiments, any table associated with a database in any information system may be identified as the table to be identified, and this application embodiment does not specifically limit this.

[0074] 102: Input multiple fields to be identified into the recognition model to perform core field recognition and obtain the recognition results of each field to be identified.

[0075] The recognition model is trained and validated using labeled core fields.

[0076] In some embodiments, core fields can be fields strongly relevant to the current production process. That is, core fields can be fields that will be used in the current production process. Annotated core fields can be obtained in advance through the following methods:

[0077] First, obtain the frequency of occurrence of each field in the information system.

[0078] Then, fields that appear more frequently than a preset frequency are identified as candidate fields.

[0079] For example, candidate fields may include "steel mill", "manufacturer", "factory", "model", "product model", etc. In this way, through initial screening, the selection range of core fields can be defined from all fields.

[0080] Specifically, the preset frequency can be set according to needs and actual conditions, and this application embodiment does not impose specific limitations on it.

[0081] Next, the core fields are determined from the candidate fields.

[0082] When determining core fields, methods such as manual annotation or semantic analysis can be used, and this application embodiment does not limit this approach. In this way, fields within the selection range are further filtered to obtain core fields that are strongly relevant to current production.

[0083] Finally, the field names of the core fields are labeled.

[0084] For example, "steel mill," "manufacturing plant," and "factory" are all labeled as "factory," and "model" and "product model" are all labeled as "product type." In this way, core fields with the same meaning but different field names can have the same label.

[0085] In some embodiments, the recognition model may be configured to perform the following steps:

[0086] Step 1: Extract the features of each field to be identified.

[0087] For example, natural language processing techniques can be used to scan the field names and descriptions of each field to be identified to obtain the features of each field. Here, both the field name and description belong to the information of the field to be identified.

[0088] Step two: Compare the features of the field to be identified with the features of each core field obtained in advance to determine the similarity between the field to be identified and each core field.

[0089] For example, the k-means algorithm can be used to cluster the features of the field to be identified and the features of each core field to obtain the clustering results of the field to be identified. If the clustering results of the field to be identified contain core fields, the similarity between the field to be identified and the contained core fields is obtained. If the clustering results of the field to be identified do not contain core fields, it means that the similarity between the field to be identified and the preset core fields in the identification model is very small, and there is no need to calculate the similarity again, thus saving processing time.

[0090] Step 3: If there is a core field whose similarity to the field to be identified is greater than a preset threshold, then the field to be identified is determined to be a core field.

[0091] Step 4: If there is no core field whose similarity to the field to be identified is greater than the preset threshold, then the field to be identified is determined to be a non-core field.

[0092] In this way, the field to be identified can be quickly matched with the core fields that have been labeled in the recognition model, resulting in very high recognition efficiency.

[0093] It should be noted that if there are a large number of fields to be identified, they can be divided into multiple threads for parallel recognition. For example, the table to be identified can be divided into multiple groups, and each thread can scan the fields to be identified in one group of tables and use the corresponding recognition model for recognition. This application does not specifically limit this.

[0094] 103: If the recognition result of the field to be recognized indicates that the field to be recognized is a core field, then obtain the labeled field name of the core field corresponding to the field to be recognized.

[0095] For example, if the field to be identified is "manufacturing plant" and the corresponding core field identified is "branch plant", then the label field name "factory" of the core field is obtained.

[0096] 104: If there are multiple label field names for the core field corresponding to the field to be identified, then based on the information of the table to be identified, determine the label field name of the target core field that has the highest matching degree with the field to be identified from the multiple label field names of the core fields corresponding to the field to be identified.

[0097] The information of the table to be identified may include at least one of the following: the field values ​​of each field to be identified, the comments for each field to be identified, the table name, and the comments for the table. Additionally, other related information can be set as needed.

[0098] For example, if the field to be identified is "type", and the corresponding core field has two labeled field names, namely "material type" and "equipment type", then based on the table name of the table to be identified or the field name and field value of other fields to be identified in the table to be identified, it is determined that the table to be identified contains equipment-related information. Therefore, "equipment type" is determined as the labeled field name of the target core field with the highest matching degree with the field to be identified, "type".

[0099] Alternatively, AI (Artificial Intelligence) modeling can be used to associate the relationships between table names, field names, field values, and labeled field names. The AI ​​model can automatically determine the target core field with the highest matching degree from the field names of multiple core fields, which can further improve the recognition efficiency.

[0100] 105: Rename the field name of the field to be identified to the label field name of the target core field.

[0101] For example, if the field name "Type" in the table to be identified is identified and the target core field with the highest matching degree is identified, the label field name "Type" of the field to be identified is renamed to "Device Type".

[0102] It should be noted that the renaming of each field to be identified can be performed simultaneously or separately, and this application embodiment does not limit this.

[0103] In addition, if the recognition result of the field to be recognized indicates that the field to be recognized is a non-core field, the label field name of the core field corresponding to the field to be recognized can be set to empty, or other identifiers.

[0104] Figure 2 This is a schematic diagram illustrating the workflow of a core field identification method provided in other embodiments of this application. (Refer to...) Figure 2 As shown, in addition to steps 101 to 105 above, the core field identification method provided in other embodiments of this application may also include the following steps 106-108, wherein step 106 can be executed in parallel with step 104, and steps 107 and 108 can be executed after steps 105 and 106.

[0105] 106: If the label field name of the core field corresponding to the field to be identified is the same, then rename the field name of the field to be identified to the label field name of the core field.

[0106] 107: Based on the recognition results of each field to be recognized, identify the erroneous fields that do not conform to the core fields and the missing fields that fail to recognize the core fields.

[0107] For example, from the fields to be identified that are indicated as non-core fields in the identification results, the missing fields whose meanings match the core fields can be filtered out. The occurrence of missing fields may be due to the absence of the corresponding core fields or incomplete feature extraction of the corresponding core fields.

[0108] For another example, from the various fields to be identified that are indicated as core fields by the identification results, fields that do not conform to the core fields can be filtered out. The occurrence of incorrectly identified fields may be due to inaccurate extraction of the corresponding core field features.

[0109] 108: Based on the identified erroneous fields, missed fields, and pre-acquired core fields, retrain the identification model and update the parameters of the identification model.

[0110] For example, features of core fields can be supplemented based on missing fields, or the corresponding core fields can be added to the recognition model. Features of core fields are then corrected based on the incorrectly identified fields. The recognition model is then retrained using all the supplemented and improved labeled core fields, making the model more closely match actual production conditions and more effectively improving recognition accuracy.

[0111] It should be noted that steps 107 and 108 can be performed simultaneously with the use of the recognition model, or they can be executed periodically.

[0112] Thus, the core field identification method provided in this application uses an identification model to determine whether each field to be identified is a core field, and uses the field name of the core field corresponding to the field to be identified to rename the field to be identified. This not only enables rapid identification of the field to be identified with high efficiency, but also allows for standardized and unified naming of each field to be identified since the core fields have been pre-labeled, thereby improving data utilization efficiency and laying the foundation for cross-departmental data linkage analysis.

[0113] The following are system embodiments of this application, which can be used to execute the method embodiments of this application. For details not disclosed in the system embodiments of this application, please refer to the method embodiments of this application.

[0114] Figure 3 This is a schematic diagram of the structure of a core field recognition system provided in an embodiment of this application. For example... Figure 3 As shown, the system provided in this application embodiment has the function of implementing the above-described core field identification method. This function can be implemented in hardware or by hardware executing corresponding software. The system may include: a field to be identified acquisition unit 301, a core field identification unit 302, a core field acquisition unit 303, a target core field determination unit 304, and a renaming unit 305. Wherein:

[0115] The field to be identified acquisition unit 301 is used to acquire the table to be identified, which includes multiple fields to be identified.

[0116] The core field recognition unit 302 is used to input multiple fields to be recognized into the recognition model for core field recognition, and obtain the recognition results of each field to be recognized. The recognition model is trained and validated using the labeled core fields.

[0117] The core field acquisition unit 303 is used to acquire the labeled field name of the core field corresponding to the field to be identified if the recognition result of the field to be identified indicates that the field to be identified is a core field.

[0118] The target core field determination unit 304 is used to determine the label field name of the target core field with the highest matching degree from the label field names of the multiple core fields corresponding to the field to be identified, based on the information of the table to be identified.

[0119] The renaming unit 305 is used to rename the field name of the field to be identified to the label field name of the target core field.

[0120] In one possible implementation, the renaming unit 305 is also used for:

[0121] If the label field name of the core field corresponding to the field to be identified is the same, then the field name of the field to be identified will be renamed to the label field name of the core field.

[0122] In one possible implementation, the core field identification unit 302 is specifically used for:

[0123] Extract features from each field to be identified.

[0124] The features of the field to be identified are compared with the features of each core field obtained in advance to determine the similarity between the field to be identified and each core field.

[0125] If there is a core field whose similarity to the field to be identified is greater than a preset threshold, then the field to be identified is determined to be a core field.

[0126] In one possible implementation, the core field identification unit 302 is specifically used for:

[0127] By scanning the field names and descriptions of each field to be identified using natural language processing technology, the features of each field to be identified are obtained.

[0128] In one possible implementation, the core field identification unit 302 is specifically used for:

[0129] The k-means algorithm is used to cluster the features of the field to be identified and the features of each core field to obtain the clustering results of the field to be identified.

[0130] If the clustering results of the field to be identified contain core fields, then obtain the similarity between the field to be identified and the core fields it contains.

[0131] In one possible implementation, the information of the table to be identified includes at least one of the following: the field values ​​of each field to be identified, the comments of each field to be identified, the table name of the table to be identified, and the comments of the table to be identified.

[0132] In one feasible approach, the core fields are those strongly relevant to current production, and the labeled core fields are obtained in advance using the following method:

[0133] Obtain the frequency of occurrence of each field in the information system.

[0134] Fields that appear more frequently than a preset frequency are identified as candidate fields.

[0135] The core field is determined from the candidate fields.

[0136] Label the field names of the core fields.

[0137] In one possible implementation, the system further includes a model update unit, which is used for:

[0138] Based on the recognition results of each field to be recognized, identify the erroneous fields that do not conform to the core fields and the missing fields that fail to recognize the core fields.

[0139] Based on the identified erroneous fields, missed fields, and pre-acquired core fields, the recognition model is retrained and its parameters are updated.

[0140] Thus, the core field identification device provided in this application uses an identification model to determine whether each field to be identified is a core field, and uses the field name of the core field corresponding to the field to be identified to rename the field to be identified. This not only enables rapid identification of the field to be identified with high efficiency, but also allows for standardized and unified naming of each field to be identified since the core fields have been pre-labeled, thereby improving data utilization efficiency and laying the foundation for cross-departmental data linkage analysis.

[0141] The present application has been described in detail above with reference to specific embodiments and exemplary examples; however, these descriptions should not be construed as limiting the present application. Those skilled in the art will understand that various equivalent substitutions, modifications, or improvements can be made to the technical solutions and implementation methods of the present application without departing from the spirit and scope of the present application, and all such modifications and improvements fall within the scope of the present application. The scope of protection of the present application is determined by the appended claims.

Claims

1. A core field identification method, characterized in that, include: Obtain the table to be identified, which includes multiple fields to be identified; Multiple fields to be identified are input into the recognition model for core field recognition to obtain the recognition results of each field to be identified. The recognition model is trained and validated using the labeled core fields. If the recognition result of the field to be identified indicates that the field to be identified is a core field, then obtain the labeled field name of the core field corresponding to the field to be identified; If there are multiple labeled field names for the core field corresponding to the field to be identified, then based on the information of the table to be identified, the labeled field name of the target core field with the highest matching degree with the field to be identified is determined from the labeled field names of the multiple core fields corresponding to the field to be identified. Rename the field name of the field to be identified to the labeled field name of the target core field; The core fields are those strongly related to current production, and the labeled core fields are obtained in advance through the following methods: Obtain the frequency of occurrence of each field in the information system; Fields that appear more frequently than a preset frequency are identified as candidate fields; Identify the core fields from among the candidate fields; The field names of the core fields are labeled; The core field recognition method also includes: Based on the recognition results of each of the fields to be identified, identify the incorrect fields that do not conform to the core fields and the missing fields that did not identify the core fields from each of the fields to be identified; Based on the identified error fields, the missed identification fields, and the pre-acquired core fields, the identification model is retrained and the parameters of the identification model are updated.

2. The method according to claim 1, characterized in that, Also includes: If the label field name of the core field corresponding to the field to be identified is the same, then the field name of the field to be identified is renamed to the label field name of the core field.

3. The method according to claim 1, characterized in that, The step of inputting multiple fields to be identified into a recognition model for core field recognition to obtain recognition results for each field to be identified includes: Extract the features of each of the fields to be identified; The features of the field to be identified are compared with the features of each core field obtained in advance to determine the similarity between the field to be identified and each core field. If there is a core field whose similarity to the field to be identified is greater than a preset threshold, then the field to be identified is determined to be a core field.

4. The method according to claim 3, characterized in that, The extraction of features from each of the fields to be identified includes: By scanning the field names and field descriptions of each field to be identified using natural language processing technology, the features of each field to be identified are obtained.

5. The method according to claim 3, characterized in that, The features of the field to be identified are compared with the features of each pre-acquired core field to determine the similarity between the field to be identified and each core field, including: The k-means algorithm is used to cluster the features of the field to be identified and the features of each core field to obtain the clustering results of the field to be identified. If the clustering result of the field to be identified contains a core field, then the similarity between the field to be identified and the core field is obtained.

6. The method according to claim 1, characterized in that, The information of the table to be identified includes at least one of the following: the field value of each field to be identified, the comment of each field to be identified, the table name of the table to be identified, and the comment of the table to be identified.

7. A core field recognition system, characterized in that, include: A field to be identified acquisition unit is used to acquire a table to be identified, wherein the table to be identified includes multiple fields to be identified; The core field recognition unit is used to input multiple fields to be recognized into the recognition model to perform core field recognition and obtain the recognition results of each field to be recognized. The recognition model is trained and verified using the labeled core fields. The core field acquisition unit is used to acquire the labeled field name of the core field corresponding to the field to be identified if the recognition result of the field to be identified indicates that the field to be identified is a core field. The target core field determination unit is used to determine the label field name of the target core field that has the highest matching degree with the target core field from the label field names of the multiple core fields corresponding to the target field, based on the information of the table to be identified, if there are multiple field names of the core field corresponding to the target field. The renaming unit is used to rename the field name of the field to be identified to the labeled field name of the target core field; The core fields are those strongly related to current production, and the labeled core fields are obtained in advance through the following methods: Obtain the frequency of occurrence of each field in the information system; Fields that appear more frequently than a preset frequency are identified as candidate fields; Identify the core fields from among the candidate fields; The field names of the core fields are labeled; The core field recognition system also includes a model update unit, which is used for: Based on the recognition results of each of the fields to be identified, identify the incorrect fields that do not conform to the core fields and the missing fields that did not identify the core fields from each of the fields to be identified; Based on the identified error fields, the missed identification fields, and the pre-acquired core fields, the identification model is retrained and the parameters of the identification model are updated.

8. The system according to claim 7, characterized in that, The renaming unit is also used for: If the label field name of the core field corresponding to the field to be identified is the same, then the field name of the field to be identified is renamed to the label field name of the core field.

Images