It is always super hyper important to know how to deal with NaN values.
It can have a huge impact on your algorithm performance, especially in Machine Learning.
You will always need to think twice about how to deal with those NaN values.
The Pandas library has few methods that you can use to deal with those NaN values, such as:
- Backward fill
- Forward fill
- Interpolate value
This method will take the next available value as present value if exists.
This method will take the previous available value as present value if exists.
This method will try to estimate what is the value in-between two observations.
This method will replace NaN values with a specific value.
Here you are, all those articles will teach you how to apply a specific method to deal with NaN values.