Combining-ILP-and-Maximum-Entropy

This repo contains my undergraduate thesis work where I tried to combine Inductive logic programming with Maximum Entropy.

Real life data does not come with good features. However, machine learning models work well only with good features. Under Dr. Ashwin Srinivasan's supervision, I experimented with Aleph ILP System to induce features. This was done in Prolog.

These features were then used to contruct a MaxEnt model. The model was compared on a set of data sets with their existing statistics.

The report is in the pdf.