Machine learning feedback loop for maximizing efficiency in transaction flow

摘要

A computer system includes a predictor engine, an extractor engine, and an acquirer engine. The predictor engine obtains resource data and stack ranks the data based on an expenditure amount each resource consumed. Resources are categorized into a tree structure. The predictor engine compiles a list identifying predicted performance enhancements and generates a notification detailing the list. In response determining that the predicted performance enhancements satisfy a threshold, the extractor engine generates an inventory of resources and rearranges the inventory based on the categories. The inventory is used to generate a baseline contrasting performance requirements with performance usages. The acquirer engine selects a resource to be used for a category based on the baseline, a list of potential resources, and the performance requirements. The acquirer engine generates or modifies an executable document. The computer system modifies these operations based on feedback to iteratively improve this transactional flow.