top of page
Feature extraction is usually the pre-requisite step to good machine learning, but can be daunting for novices and those brand new to the idea of machine learning. It can also be a challenge for niche or overly complicated structured datasets. We will talk about pyLZJD, which implements a compression-based approach to machine learning. This approach allows pyLZJD to obtain reasonable results on many tasks by simply providing raw byte data, with no feature engineering required. For some tasks with complex structure, like malware similarity, pyLZJD may even be state of the art.
bottom of page