Propheticus - Generalizable Machine Learning Framework
Propheticus is a (generalizable) Machine Learning framework that includes functionalities for all the steps in a Machine Learning work, from data analysis and preprocessing, to model assessment and comparison. It was created to overcome the limitations of existing tools towards research in the dependability area, which requires it to be flexible and adaptable to fit the needs of the users. Propheticus can be applied to a variety of problems (e.g. error detection, failure prediction, intrusion detection) to create models whose predictions can then be used to develop and deploy more dependable systems.
The version available in this repository is the same submitted to the IEEE 30th International Symposium on Software Reliability Engineering (ISSRE). The framework is still under development to broaden its scope and functionalities, as well as to improve its structure, coding standards, usability, and documentation.