yzhao062/pyod

specifying categorical features in Python Outlier Detection (PyOD)

Open

#21 opened on Sep 14, 2018

View on GitHub
 (5 comments) (0 reactions) (1 assignee)Python (7,762 stars) (1,308 forks)batch import
enhancementhelp wanted

Description

How to specify the categorical features in PyOD when using Histogram-based Outlier Detection (HBOS) for anomaly detection ? I've read that HBOS can be used for anomaly detection when there are categorical features involved. I found it's Python implementation here: https://pyod.readthedocs.io/en/latest/pyod.models.html#module-pyod.models.hbos But I can't figure out how should I pass the position or list of names of categorical features of my dataset while training the model. The code I've tried:

clf = HBOS(n_bins=10, alpha=0.1, tol=0.5, contamination=0.1)
clf.fit(train_df)
train_pred = clf.labels_

There is no parameter to mention categorical features while training.

Contributor guide

Tech stack
python
Domain
datamachine learning
Issue type
feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
half day
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
clear
Prerequisites
Basic PythonFamiliarity with PyODUnderstanding of HBOS
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
40
Research direction
Investigate the HBOS implementation in pyod/models/hbos.py to understand current handling of features. Determine if categorical features are already handled implicitly or if a new parameter is needed. Review comments on this issue for any proposed solutions or workarounds. Implement a parameter 'categorical features' that accepts indices or column names, and adjust binning accordingly.