System and method for predictive structuring of electronic data

Patent metadata

Field	Value
Publication number	US-11892989-B2
Application number	US-202217705448-A
Country	US
Kind code	B2
Filing date	Mar 28, 2022
Priority date	Mar 28, 2022
Publication date	Feb 6, 2024
Grant date	Feb 6, 2024

Abstract

Embodiments of the invention are directed to a system, method, or computer program product for an approach to predictive structuring of electronic data. The system uses data mining and data clustering techniques using classification models to organize feature column groups comprising feature columns. The system identifies and flags feature column groups and/or feature columns based on regulatory data standards provided by regulatory bodies. Thereafter, data objects are imported into the system and prediction algorithms are implemented to characterize the feature columns containing the data objects.

First claim

What is claimed is:

1. A system for organizing data objects, the system comprising:
a memory device with computer-readable program code stored thereon;
a communication device;
a processing device operatively coupled to the memory device and the communication device, wherein the processing device is configured to execute the computer-readable program code to:
receive entity data tables, wherein the entity data tables comprise data objects stored in a memory device of an entity system;
receive regulatory standard documentation comprising data handling guidelines identified from keywords in text of the regulatory standard documentation, wherein the regulatory standard documentation is a pdf file;
mine metadata from the entity data tables and mine the data handling guidelines from the regulatory standard documentation;
perform clustering on mined data using a K-means algorithm to determine a common grouping of the mined data to prepare for identification of a feature column group, wherein the mined data comprises the mined data handling guidelines and the mined metadata, and wherein the feature column group comprises one or more feature columns determined by the K-means algorithm to relate to other feature columns within the feature column group based on the metadata and data handling guidelines of the data objects within each of the feature columns;
import unknown data objects to a cell of a column;
identify the feature column of the column using a machine learning model trained using training data objects with known feature columns and known feature columns groups;
identify the feature column group based on data type and sensitivity; and
improve a confidence interval of the identified feature column of the feature column group by subjecting the data objects to the machine learning model multiple times.

2. The system of claim 1 , wherein the processing device is further configured to execute the computer-readable program code to:
assign a related target key to data objects extending across multiple feature column groups;
store the related target key in a relationship key table;
append the relationship key table to each of the feature column groups; and
output predictively structured data in one or more visual formats.

3. The system of claim 1 , wherein performing clustering on the mined entity data tables is performed iteratively.

4. The system of claim 1 , wherein the machine learning model further comprises classifying and identifying the data objects or the one or more feature columns as numeric or non-numeric, and wherein a Naïve Bayes algorithm is used for non-numeric, and a decision tree algorithm is used for numeric.

5. The system of claim 1 , wherein identifying the feature column comprises determining a name of the feature column.

6. The system of claim 1 , wherein the processing device is further configured to execute the computer-readable program code to:
flag the feature column group based on the data type and the sensitivity.

7. The system of claim 1 , wherein the processing device is further configured to execute the computer-readable program code to:
flag the feature column based on the data type and the sensitivity.

8. A computer program product for organizing data objects, the computer program product comprising at least one non-transitory computer-readable medium having computer-readable program code portions embodied therein, the computer-readable program code portions comprising:
receiving entity data tables, wherein the entity data tables comprise data objects stored in a memory device of an entity system;
receiving regulatory standard documentation comprising data handling guidelines in text identified from keywords in text of the regulatory standard documentation, wherein the regulatory standard documentation is a pdf file;
mining metadata from the entity data tables and mining the data handling guidelines from the regulatory standard documentation;
performing clustering on mined data using a K-means algorithm to determine a common grouping of the mined data to prepare for identification of a feature column group, wherein the mined data comprises the mined data handling guidelines and the mined metadata, and wherein the feature column group comprises one or more feature columns determined by the K-means algorithm to relate to other feature columns within the feature column group based on the metadata and data handling guidelines of the data objects within each of the feature columns;
importing unknown data objects to a cell of a column;
identifying the feature column of the column using a machine learning model trained using training data objects with known feature columns and known feature columns groups;
identifying the feature column group based on data type and sensitivity; and
improving a confidence interval of the identified feature column of the feature column group by subjecting the data objects to the machine learning model multiple times.

9. The computer program product of claim 8 , the computer-readable program code portion further comprising:
assigning a related target key to data objects extending across multiple feature column groups;
storing the related target key in a relationship key table;
appending the relationship key table to each of the feature column groups; and
outputting predictively structured data in one or more visual formats.

10. The computer program product of claim 8 , wherein performing clustering on the mined entity data tables is performed iteratively.

11. The computer program product of claim 8 , wherein the machine learning model further comprises classifying and identifying the data objects or the one or more feature columns as numeric or non-numeric, and wherein a Naïve Bayes algorithm is used for non-numeric, and a decision tree algorithm is used for numeric.

12. A computer-implemented method for organizing data objects, the method comprising:
providing a computing system comprising a computer processing device and a non-transitory computer readable medium, where the non-transitory computer readable medium comprises configured computer program instruction code, such that when said computer program instruction code is operated by said computer processing device, said computer processing device performs the following operations:
receiving entity data tables, wherein the entity data tables comprise data objects stored in a memory device of an entity system;
receiving regulatory standard documentation comprising data handling guidelines in text identified from keywords in text of the regulatory standard documentation, wherein the regulatory standard documentation is a pdf file;
mining metadata from the entity data tables and mining the data handling guidelines from the regulatory standard documentation;
performing clustering on mined data using a K-means algorithm to determine a common grouping of the mined data to prepare for identification of a feature column group, wherein the mined data comprises the mined data handling guidelines and the mined metadata, and wherein the feature column group comprises one or more feature columns determined by the K-means algorithm to relate to other feature columns within the feature column group based on the metadata and data handling guidelines of the data objects within each of the feature columns;

importing unknown data objects to a cell of a column;
identifying the feature column of the column using a machine learning model trained using training data objects with known feature columns and known feature columns groups;
identifying the feature column group based on data type and sensitivity; and
improving a confidence interval of the

Assignees

• Bank Of America

Inventors

• Jain Deepak
• Gavhane Bhakti Sanjay
• Sandhu Simranjeet Singh
• Sharma Aditya

Classifications

• G06F16/213 Primary

  with details for schema evolution support · CPC title

• G06F16/2465 Primary

  Query processing support for facilitating data mining operations in structured databases · CPC title

• G06F16/258

  Data format conversion from or to a database · CPC title

• G06F16/285

  Clustering or classification · CPC title

• G06F16/254

  Extract, transform and load [ETL] procedures, e.g. ETL data flows in data warehouses · CPC title