Skip to content

HiteshKongadi/Advanced-Regression

Repository files navigation

Advanced Regression using Ridge and Lasso Regularization

The model is expected to help the management understand the key factors affecting the prices of house.

Table of Contents

General Information

  • A US-based Housing company wants to understand the variables affecting the sale of houses.
  • The company wants to enter into Australian Housing market.
  • Trying to understand the contributing factors which can give us insights in to the Australian housing market when the company starts to operate in Australian market.
  • Data set used is the Housing market data from US which has the features which may or may not be affect the prices of the houses.(train.csv)

Conclusions

  • Ridge and Lasso Regression Model are built with optimum alpha calculated in GridSearchCV method. Optimum alpha = 9.0 for ridge and 0.0001 for lasso model.
  • Model evaluation is done with R2 score and Root Mean Square Error.
  • Lasso Regression is chosen as final model for having slightly better R-square value on test data.
  • Out of 50 features in the final model, top 10 features in order of descending importance are ['1stFlrSF', '2ndFlrSF', 'OverallQual', 'OverallCond', 'SaleCondition_Partial', 'LotArea', 'BsmtFinSF1','SaleCondition_Normal', 'MSZoning_RL', 'Neighborhood_Somerst']

Technologies Used

Python

  • library - Pandas
  • library - NumPy
  • library - matplotlib
  • library - Seaborn
  • library - sklearn
  • library - statsmodel

Acknowledgements

Give credit here.

Contact

Created by [@HiteshKongadi] - feel free to contact me!

About

Ridge and Lasso regression models Assignment

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published