All Courses
AI Career Planner
1:1 Coaching/Mentoring
Job Board
SQL
HTML/CSS/JS
Coding
Settings
LogoutBlatant outlier removal on grounds of mean and standard deviation can lead to removal of data points that might be necessary for the analysis, especially if the sample size is small. Also outlier removal many a times is not a one time process. Consider removing a first set of obvious outliers using a combination of stats and domain knowledge. For e.g. weight of a 30 year old as 10 kg is probably an outlier, whereas weight of a 2 year as 10 kg is definitely not.
Explore the data after removing outliers, and look out for more outliers. Continue the process till the point outliers data points are sufficiently smaller as compared to the sample size.