US 2023/0419130 A1Published application

Machine learning pipeline with data quality framework

AssigneeOracle Financial Services Software Limited

InventorsSharoon SAXENA, Veresh JAIN, Rahul YADAV

Life of this patent

Estimating the term…

Loading patent…

Abstract

Systems, methods, and other embodiments associated with a data quality framework for a machine learning pipeline are described. In one embodiment, a method includes receiving an input dataset prior to training a machine learning model with the input dataset. The input dataset, having data records in column tabular form, is analyzed to determine a data quality of the input dataset prior to training the machine learning model. A data quality score is generated that represents an overall quality of the input dataset, where the data quality score is generated based on a combined ensemble of at least two factors selected from (i) an abundance factor, (ii) a completeness factor, and (iii) a dimension efficiency factor. Based on the data quality score, the input dataset is either permitted or prohibited from continuing in the machine learning pipeline, and corrective actions may be performed.

Claims20

Dependent claims

A computing system, comprising:

at least one processor connected to at least one memory;
a machine learning pipeline configured to receive an input dataset and train a machine learning model using the input dataset;
a data quality framework operably connected to the machine learning pipeline and configured to determine a data quality of the input dataset prior to training the machine learning model;
wherein the data quality framework includes instructions stored on a non-transitory computer readable that when executed by at least the processor cause the processor to:
generate a data quality score that represents an overall quality of the input dataset;
wherein the data quality score is generated based on a harmonic sum of (i) an abundance factor of the input dataset, (ii) a completeness factor of the input dataset, and (iii) a dimension efficiency factor of the input dataset;
in response to the data quality score not meeting a quality threshold, prohibit the input dataset from continuing in the machine learning pipeline;
and
generate a notification of poor data quality and generate one or more corrective actions for improving the quality of the input dataset;
and
in response to the data quality score meeting the quality threshold, allow the input dataset to continue in the machine learning pipeline and be used to train the machine learning model.

8 more elements

A computer-implemented method, the method comprising:

receiving an input dataset prior to training a machine learning model with the input dataset in a machine learning pipeline;
analyzing the input dataset having data records in column tabular form, to determine a data quality of the input dataset prior to training the machine learning model;
generating a data quality score that represents an overall quality of the input dataset;
wherein the data quality score is generated based on a combined ensemble of at least two factors selected from:
(i) an abundance factor of the input dataset, (ii) a completeness factor of the input dataset, and (iii) a dimension efficiency factor of the input dataset;
in response to the data quality score not meeting a quality threshold:
prohibiting the input dataset from continuing in the machine learning pipeline;
generating and displaying a notification of poor data quality of the input dataset;
and
generating one or more corrective actions for improving the quality of the input dataset;
and
in response to the data quality score meeting the quality threshold, allowing the input dataset to continue in a machine learning pipeline and be used to train the machine learning model.

9 more elements

A non-transitory computer-readable medium that includes stored thereon computer-executable instructions that when executed by at least a processor of a computer cause the computer to:

receive an input dataset prior to training a machine learning model with the input dataset in a machine learning pipeline;
analyze the input dataset having data records in column tabular form, to determine a data quality of the input dataset prior to training the machine learning model;
generate a data quality score that represents an overall quality of the input dataset based on the analyzing;
wherein the data quality score is generated based on a harmonic sum of at least two factors selected from:
(i) an abundance factor of the input dataset, (ii) a completeness factor of the input dataset, and (iii) a dimension efficiency factor of the input dataset;
in response to the data quality score not meeting a quality threshold:
prohibiting the input dataset from continuing in the machine learning pipeline;
and
perform corrective actions to create a modified input dataset based on the at least two factors including:
(i) identifying and removing redundant columns in the input dataset, and (ii) identifying and removing columns with an amount of missing values that exceeds a threshold;
and
repeat analyzing the modified input dataset and generate an updated data quality score.

9 more elements

Dependencies are inferred from the claim text.