statsmodels|sklearn|matplotlib|pandas|numpy

ols|LinearRegression|LeaveOneOut|Kfold|train_test_split

Task : build a simple linear regression model to predict salary using experience

Summary : I build a simple linear regression model using :

• ols method
• LinearRegression

I used ols to understand some basic terms like R-squared, adj. R-squared, AIC, significance of p-value. Further I build model using LinearRegression. The techniques I used in building model are - LeaveOneOut, Kfold, train_test_split to compare all these with the same dataset.

Conclusion :

• Model build using ols has R-squared = 95% which is good and LinearRegression have accuracy approximately in the line of 90's.
• All the three techniues - LeaveOneOut, Kfold, train_test_split give approximately same accuracy except LeaveOneOut (60%) and this might be due to very small dataset I choose to build model.