We chose categorical-crossentropy loss function because we got maximum accuracy using it compared to other loss function models.
- Initially we clean the data.
- Missing data was replaced by mean of group for continous and mode of group for categorical variables.
- we replaced categorical variables by corresponding data value basically we converted to string data values to integer data values
- we standardized the data using standard scalar function
- Then we split the data for training and testing
- Then we build the model