This is my implementation of the Isolation Forest anomaly detection algorithm from a paper by Fei Tony Liu, Kai Ming Ting, and Zhi-Hua Zhou. The code is adapted from a project I did for a course in Algorithms and Data Structures taught by Terence Parr as part of the University of San Francisco Master of Science in Data Science program.
The original algorithm is pretty slow in Python. Instead of randomly splitting the data as described by Liu, Ting, and Zhou, I randomly partitioned the data 5 times and then chose the partition that created the biggest difference in size between the two parts. This improved the efficiency of the algorithm by causing anomalies to be isolated sooner.
I used a breast cancer dataset in Scikit-learn and this credit card fraud dataset from Kaggle.