Project 1
- How did your model fare?
- Not great, the loss for most epochs varied between 70 and 90, though the scatter plot had a visible trend.
- In your estimation is there a particular variable that may improve model performance?
- Zipcode appears to be one of the more important variables. In the Washington DC data I looked at most expensive homes were in either 20007 or 20008, while the less expensive homes were often in 20019 or 20020.
- Which of the predictions were the most accurate? In which percentile do these most accurate predictions reside? Did your model trend towards over or under predicting home values?
- Though none of the predictions were very accurate, the ones for cheaper houses were closer, around the 25th percentile. It overpredicted most of the values.
- Which feature appears to be the most significant predictor?
- Number of bedrooms appears to be the most important factor in home price.