Method and apparatus for dataset evaluation and database management

A method for processing a plurality of documents associated with a transaction, said method comprising the steps of:

• uploading a first plurality of data values associated with said transaction;
• uploading digital images of a plurality of documents associated with said transaction, wherein said plurality of documents include a second plurality of data values embedded therein;
• providing modified content of said digital images in which said second plurality of data values that are numerical digit values and that were embedded in said plurality of documents are omitted;
• providing access to said modified content, and responsive to said access, receiving from a first machine learning engine, document identification of said plurality of documents;
• providing access to said digital images with said second plurality of data values to a second machine learning engine, and after providing said access, receiving from said second machine learning engine, extraction of said second plurality of data values from said digital images along with category identification of respective document categories for each of said second plurality of data values that are extracted from said digital images;
• enabling comparison and confirming accuracy of said first plurality of data values and said second plurality of data values by displaying said first plurality of data values with said second plurality of data values as being associated with each other, respectively, based on said category identification, and enabling correction if based on said displaying, association between any of said first plurality of data values and said second plurality of data values is missing or incorrect;
• providing feedback regarding correctness of said document identification based on said displaying to said first machine learning engine, for retraining said first machine learning engine to increase accuracy of said document identification of said documents;
• providing feedback regarding correctness of extraction of said second plurality of data values from said digital images or said category identification to said second machine learning engine for retraining said second machine learning engine to increase accuracy of said extraction or said category identification.

A non-transitory computer readable medium having stored thereon instructions for operating a document evaluation system, the instructions when executed by at least one processor cause said at least one processor to implement the steps of:

• uploading a first plurality of data values associated with said transaction;
• uploading digital images of a plurality of documents associated with said transaction, wherein said plurality of documents include a second plurality of data values embedded therein;
• providing modified content of said digital images in which said second plurality of data values that are numerical digit values and that were embedded in said plurality of documents are omitted;
• providing access to said modified content, and responsive to said access, receiving from a first machine learning engine, document identification of said plurality of documents;
• providing access to said digital images with said second plurality of data values to a second machine learning engine, and after providing said access, receiving from said second machine learning engine, extraction of said second plurality of data values from said digital images along with category identification of respective document categories for each of said second plurality of data values that are extracted from said digital images;
• enabling comparison and confirmation of accuracy of said first plurality of data values and said second plurality of data values by displaying said first plurality of data values with said second plurality of data values as being associated with each other, respectively, based on said category identification, and enabling correction if based on said displaying, association between any of said first plurality of data values and said second plurality of data values is missing or incorrect;
• providing feedback regarding correctness of said document identification based on said displaying to said first machine learning engine, for retraining said first machine learning engine to increase accuracy of said document identification of said documents;
• providing feedback regarding correctness of extraction of said second plurality of data values from said digital images or said category identification to said second machine learning engine for retraining said second machine learning engine to increase accuracy of said extraction or said category identification.