4-16 Advanced Methods for Univariate Outlier Detection Part 1: Trimmed Mean, Median, MAD

Previous: https://youtu.be/J1wcjFngyW8 Next: https://youtu.be/X_6rnw5UM5E Playlist: https://www.youtube.com/playlist?list=PLnccJ9vccaGlC77PmUoPh2cUziN2k6yrs Univariate Outlier Detection: Trimmed Mean & MAD Method Explained with Python Code In this video, we dive into advanced methods for univariate outlier detection, focusing on two robust techniques: Trimmed Mean and Median Absolute Deviation (MAD). These methods provide more reliable results than traditional approaches like the standard Z-score, especially when dealing with datasets that include extreme values. The story begins with the Trimmed Mean method, a modification of the standard deviation method that replaces the arithmetic mean with a trimmed mean—an average calculated after removing a certain percentage of the highest and lowest values. We demonstrate how a 10% trim helps reduce the influence of extreme outliers. Next, we explore the Median and MAD (Median Absolute Deviation) approach—a robust alternative using the median instead of the mean. This technique is particularly resilient against the influence of outliers. Throughout the tutorial, we compare detection results across four methods—Z-score, IQR, Iterative, and Trimmed Mean—helping you visually assess how each method performs. While both Trimmed Mean and MAD are effective at identifying extreme high values, we observe that low-value outliers may not always be detected, revealing the strengths and limitations of each method. This is Part 1 of our advanced outlier detection series. Stay tuned for Part 2, where we will cover Robust Z-score and Winsorization techniques. 【Predictive Analytics by Machin Learning】 This course "Predictive Analytics by Machine Learning" explicates essential concepts and techniques ranging from foundational to advanced. It covers not only machine learning algorithms but also various concepts and methods for data preprocessing. This course will guide you step-by-step, equipping you with the skills to confidently apply machine learning to real-world predictive analytics. Instructor: Takuma Kimura (木村 琢磨), Ph.D. Scientist of Organizational Behavior and Analytics https://orcid.org/0000-0001-7126-188X https://www.linkedin.com/in/takuma-kimura-ba6242104/ #machinelearning #datascience #outlier #outliers #outlierdetection #trimmedmean #median #MAD