A simple linear model fitted over Boston-House Dataset having 506 datapoints.
Through this dataset, we can demonstrate -
-
Which features can best predict the target, by finding correlation between target and different features. This can be demonstrated by a heatmap.
-
The effect of outlier on predicted values. This can be demonstrated by removing outlier,and see the effect on r2score and RMSE. By removing outliers, the model become more generalised, and give better predictions.
DEPENDENCIES :-
- Pandas
- Matplotlib
- Seaborn
- Scikit-Learn