Zomato Interview Questions | Outlier Vaues

Question

Give an example of outlier values and how can they be treated? Do having more outlier values a good thing or a bad thing?

in progress 0
Dhruv2301 4 years 2 Answers 850 views Great Grand Master 1

Answers ( 2 )

  1. Priyabrata017
    1
    Reply

    Suppose, we have the speed of car every 6 minutes from 10 A.M to 11 A.M given by

    20,40,30,110,50,30,55,20,15,40.
    While calculating the average speed of the car, the speed 110 will be outlier.

    We can remove the outlier if we are certain that we are not losing any precious data( depends on the domain knowledge).
    Else we can take log or square root of the data.

  2. juhi sharma
    1
    Reply

    Outliers are the values/data points that differ significantly from other data points.
    example –
    If we have marks of 10 students of a particular class out of 50.

    5,25,26,24,28.5,27,29,25.5,45,29

    Here mostly students got marks between 24 to 30 but we have a student who scored extremely low ie 5/50 and we have a student who got 45/50.

    so 5 and 45 both will be treated as outliers but 5 will be a negative outlier and 45 will be a positive outlier.

    Outliers in the data set can be treated either by the z-score or IQR method.

    Removing the outliers depends upon the problem statement and data set.
    If we are working on fraud detection, definitely we don’t need to remove outliers. In this case, outliers will be important.
    else we can transform the data by log transformation, square root transformation etc

Leave an answer

Browse
Browse