This blog post focuses on the optimization process and evaluation of a single component used for outlier detection. We implemented our own forest inspired solution as part of our meta learning framework. Although our usage and applied implementation are quite different, we share the (fun part of the) journey. We will limit the scope of the blog post, and avoid manipulating the important data preprocessing part. Instead we will focus on the journey that led to our unique implementation. For readability purposes, we will break this blog post to 3 different parts. To access the full white paper on Pentoma’s Generative Adversarial Model Agnostic Networks (GAMAN) and the unique technology we created, please register below.