Shobha
The best way to predict the future is to invent it
How do you assess the impact of outliers in your data and model predictions?
I assess outliers by using boxplots or calculating the Z-score to identify extreme values. Outliers can skew the results, so I either remove them, transform the data, or apply robust models like Random Forests that are less sensitive to them.For example,If the age of employees in a salary prediction model includes an outlier (e.g., a 120-year-old), it may distort the relationship, and I would consider removing or transforming that value.