Predicting Average Housing Starts
1. Figure
[Feature Importance using Linear Regression ]
2. Goal
Create a model to predict the average housing starts within 1, 2, and 4 quarters by using machine learning methods.
3. Methodology & Summary
The researchers collected data from the Federal Reserve Economic Data (FRED) online database, which included two crucial data sets: GDP and recession dates, in quarterly format extending only to the 1970s. All other data was also aggregated to the quarter level, and dates prior to the 1970s were dropped.
To account for the fact that a recession is a function of what happened in the quarters prior, the r esearchers shifted the target column indicating 1 for a recession and 0 for not a recession. This was done so that the rest of the data would point ahead to whether or not there was a recession.
The researchers created three different X matrices for each shift with 1, 2, and 4 quarters, and the data was split into training and testing sets and normalized.