Fraud & Anomaly Detection via Autoencoder & Isolation Forest | Machine Learning | Community Webinar

An overview of fraud and anomaly detection via Autoencoder and Isolation Forest. “Experts estimate federal government losses from potential fraud at nearly $150 billion” and new data shows the Federal Trade Commission received 2.8 million reports of fraud in 2021 from consumers. In this webinar, we liked to show how to fight fraud with KNIME, a free and low-code tool, that can perform fraud and anomaly detection without a single line of code nor brittle if-then rules! In the first part of this webinar, we will work with labelled data to perform classical machine-learning approaches to fraud detection such as the random forest. Then we will cover a deep learning technique, the autoencoder, to find fraudulent data points. In the second part of the webinar, we will focus on data without labels of fraudulent activity using visualizations, classical statistics, and machine learning. You will learn how easy it is to generate multiple visualizations, perform statistical analysis, and use two machine learning algorithms - Isolation Forest and DBSCAN - all to detect fraudulent activity in the free, open-source KNIME Analytics Platform. In this session you will learn: - How to identify fraud using a variety of techniques including visualizations, statistics, and machine learning - How to use machine learning and deep learning algorithms for fraud detection regardless of whether you have labelled data or not Table of Contents: 00:00 Introduction 01:29 KNIME Analytics Platform 02:51 KNIME nodes & workflow 04:24 Goals for the Session 05:29 Fraud is all around us 07:07 Potentially fraudulent data 07:38 Fraudulent data might be labelled 09:14 Decision Tree 11:53 Random Forest 16:05 Advanced: Sampling Strategies 17:58 Finding fraud through deep learning 19:59 A neural autoencoder in KNIME 20:55 Walk through how to do the same task using unlabeled data (Jinwei) 21:27 Fraud and Outlier Detection 25:01 Finding Outliers: Statistics 30:43 Demo IQR and Z-score Implementation i