Streaming Isolation Forest - A presentation for a paper accepted in PAKDD

Streaming Isolation Forest

Workshop materials for Melbourne R user group 2024 December. 1. Shiny Dashboard Development using LLMs (Shiny Assistant) How we can use LLM to build Shiny Dashboard? 2. State-of-the-art Generative AI powered Shiny Dashboard (Sidebot) What can you do in R using LLM models? 3. Hand-on session to run a chatbot in Rstudio (elmer package) a. Getting an API key, b. Install a package, c. Setup the API to work with the package d. Run a chatbot

This presentation introduces the foundational concepts of anomaly detection and offers an in-depth explanation of the Isolation Forest algorithm. Using a real-life dataset, we compare multiple anomaly detection techniques, including Local Outlier Factor and One-Class Support Vector Machine, alongside three variations of Isolation Forest. The comparison evaluates their processing times and detection performance.

This presentation introduces the basics idea of Anomaly Detection, its concepts and applications. We also compare several anomaly detectors, including Local Outlier Factor, One-class Support Vector Machine and variants of Isolation Forest in terms of processing time and detection accuracy using Area Under Receiver Operating Characteristic Curve using a real-life data set.