Sumedh Sankhe
Sharyu Deshmukh
Shivayogi Biradar
When a buyer is asked to describe their dream house, they probably would begin with the number of bedrooms, availability of white picket fence, number of floors etc. and not the height of basement ceiling or the proximity to a train station. From this project, we aim to prove that much more influences price negotiations than the features mentioned above.
To build models to predict house prices, given the Ames (Iowa) Housing dataset, with high degree of predictive accuracy
The goal of the problem is to utilize machine learning tools to make the best possible prediction of house prices. This is an interesting problem because most people will eventually buy/sell a home. This problem allows us, as data scientists, to learn more about the housing market and helps with making more informed decisions.
• Linear Regression with subset selection, Lasso and Ridge. Since we have sale price as target variable linear regression is the first method we will implement on this problem. Since we have 79 features in this dataset methods like subset selection, Lasso and Ridge may help us achieve better accuracy.
• Random Forest Regressors, Decision Trees, Neural Networks, Kernel Methods and other advanced Techniques to be implemented as and when they are taught in the class.