Predicting API endpoint descriptions from API documentation

摘要

A method includes extracting, from a labelled repository, a plurality of true endpoint descriptions associated with a plurality of web APIs and retrieving a documentation corpus associated with the plurality of web APIs. The method further includes determining a plurality of false endpoint descriptions associated with the plurality of web APIs based on the retrieved documentation corpus and the extracted plurality of true endpoint descriptions. The method further includes extracting feature information associated with the plurality of web APIs, generating a training dataset based on the extracted feature information, and obtaining a ML model based on the generated training dataset. The method further includes providing a set of inputs to the ML model and classifying the provided set of inputs as one of a true description or a false description associated with an endpoint based on a prediction result of the ML model for the provided set of inputs.