Workshop / Overview

In many fraud- (or general outlier-) detection situations, labelled data is not available. We therefore need to resort to unsupervised methods to identify points that are somehow untypical.

In this workshop, a short introduction will be given that discusses the main outlier detection methods (from the classic LOF to modern algorithms such as Isolation Forest and autoencoders) and appropriate metrics for highly imbalanced datasets.

Then, participants will be given unlabelled datasets to make predictions on. Scores will be compared on a leader board, with the emphasis on comparing techniques.

Workshop / Outcome

After the workshop, participants will:

  • Know the main algorithms for unsupervised outlier detection, and their pros and cons
  • Understand what scoring metrics may be used for highly imbalanced classification, and how these relate to business costs
  • Have gained practical experience doing outlier analysis in Python

Workshop / Difficulty

Intermediate level

Workshop / Prerequisites

  • Intermediate Python skills
  • Basic understanding of Machine Learning concepts
  • Laptop with internet access (teams of two may be formed), a Google account for colab, alternatively Docker with downloaded image or with correct Python packages installed (see instructions in the Github page).

Track / Co-organizers

Ernst Oldenhof

Data Scientist, Julius Bär

AMLD EPFL 2020 / Workshops

A Conceptual Introduction to Reinforcement Learning

With Kevin Smeyers, Katrien Van Meulder & Bram Vandendriessche

09:00-12:30 January 251ABC

Applied Machine Learning with R

With Dirk Wulff, Markus Steiner & Michael Schulte-Mecklenbeck

09:00-17:00 January 25Foyer 6

Augmenting the Web browsing experience using machine learning

With Oleksandr Paraska, Vasily Kuznetsov, Tudor Avram & Levan Tsinadze

09:00-12:30 January 253A

AMLD / Global partners