Icon

IQR for time series outlier detection

This simple workflow demonstrates as to how to discover outliers in a time series using Inter Quartile Range method. This technique fails to discover any outlier in this particular case.

URL: Outlier detection techniques https://www.kaggle.com/code/alexeslava/outlier-detection-techniques-for-time-series

Generate a sine wave. Add few noisy outliers. Discover them using IQRThis technique doe not discover all the outliers Refer: Kaggle: https://www.kaggle.com/code/alexeslava/outlier-detection-techniques-for-time-series decidewhich yare outliersbased on missing valuesin ycreate var: kdefault: 1.5decidecolor ofoutliersscatter plotoft vs ydata: t vs ysine wavewith noisesine wavewith noiseReplacey by ?, if > k * IQR Rule Engine DoubleConfiguration Color Manager Scatter Plot(JavaScript) Table Reader Scatter Plot(JavaScript) Line Plot(JavaScript) Numeric Outliers Generate a sine wave. Add few noisy outliers. Discover them using IQRThis technique doe not discover all the outliers Refer: Kaggle: https://www.kaggle.com/code/alexeslava/outlier-detection-techniques-for-time-series decidewhich yare outliersbased on missing valuesin ycreate var: kdefault: 1.5decidecolor ofoutliersscatter plotoft vs ydata: t vs ysine wavewith noisesine wavewith noiseReplacey by ?, if > k * IQR Rule Engine DoubleConfiguration Color Manager Scatter Plot(JavaScript) Table Reader Scatter Plot(JavaScript) Line Plot(JavaScript) Numeric Outliers

Nodes

Extensions

Links