HousePrices

Goal

Predict the sales price for each house, based on a set of features.

Dataset

The Ames Housing is the data set used in this project, which was compiled by Dean De Cock for data science education purposes.

What we did

Since this was one of the first kaggles we participated on we tried a little bit of everything. For this reason, this repository may have become a little messy so here's a description of the major achievements in each file:

• /Feature Engineer.py
  • Dummy variables
  • Forward Feature Selection using the R2
  • Regression models: Ridge and Lasso
• /housePrices_onehotencoder.py
  • Applying Onehot Encoding to the categorical features
  • Building linear regression models with them.
• /houseprices.py
  • Main data analysis.
  • Plotting the correlation matrix.
  • Feature selection by the correlation to the output and
  • Using the top 3 to make a non-linear regression model.
• /PCA.py
  • tried to apply PCA but the results were not ideal (will try again more thoroughly in the future)
• /src/backward_elimination.py
  • Feature selection through Backward Elimination method using the p-value
• /src/feat_eng_and_SVR.py
  • Support Vector Regressor
• /src/OHE_and_ElasticNet.py
  • Elastic Net regression model
• /src/main
  • main points in a concise and clear code

Note:

the code files in /src use the kaggle framework built to summarize the pre-processing part of the code in just a couple of lines, ideally in any Kaggle challenge. To use it for the first time you need to follow these quick steps:

1. Clone the framework repository: https://github.com/AlexGomesDS/Kaggle_project_framework
2. Add the folder /src in it to sys.path, for example, if you cloned to "C:\GIT\Kaggle_project_framework" run:

import sys
sys.path.append(r'C:\GIT\Kaggle_project_framework')