CODEWITHSUNDEEP
weka
Sudhakar Chavan
Sudhakar Chavan

Reputation: 377

Outlier treatment in Weka Explorer

Once outliers are identified in Weka, instead of removing them how do i treat the values. For example a dataset has column credit_balance which has values starting from 1 to 1000, and after applying formula for interquartile range we get the outlier records with values more than 800. So now instead of removing these records i want to replace the values 800 with mean as 300. how can we do that in Weka Explorer?

Thanks

Upvotes: 1

Views: 2985

Answers (1)

Sudhakar Chavan
Sudhakar Chavan

Reputation: 377

Crude method : Explore with below navigation. Flow > Preprocess > Filter > Choose > Filters > Unsupervised > Attribute > AddExpression – this option will create new field e.g : ifelse(a2 > 1000,200, 1) here a2 - is your attribute number.

Limitation : This will work for specific attributes only.

Upvotes: 1

Related Questions