Authors: Alex Bruening and Ryan Willett
Date of final publication: June 3rd, 2019

This repo is for the classic Kaggle competition concerning prediction of the sale prices of houses in Ames, Iowa.
Documents present in the repo are limited to exploratory Jupyter notebooks, the data work-up demonstrating pipeline usage, and an example script that demonstrates the workflow for using the aforementioned documents.

A discussion of this repository can be found at this blog post.

Versions used reflect the latest available at the time of publication.

For more information about this project feel free to contact ajb0211 "at" gmail "dot" com.

• Published blog post
• Limited expanded .gitignore for legibility
• Uploaded model training
• Deprecated old models

• Restructured folders
• Universal feature engineering file
• SVR class and parameter search
• Added example file showing how to do parameter search with an ElasticNet model

• Add ensemble model
• Base model class for ensembled models
• New model classes for:
  • ElasticNet
  • LightGBM
  • Ridge
  • GradientBoostingRegressor
• Used RobustScaler and PowerTransformer for improvements in evaluation
• Add external workup scripts for each model

• Finished light gbm model. merge from work on desktop

• More linear models.
• Began lightGBM model training

• 4 kaggle submissions of linear models
• Integration of all features
• Completed construction of linear model pipeline

• Began constructing pipelines

• Git repo init
• Moved old files from Kaggle to local