Skip to content

Latest commit

 

History

History
19 lines (16 loc) · 989 Bytes

README.md

File metadata and controls

19 lines (16 loc) · 989 Bytes

King County

Making my own scuffed Zillow

Here is my stab at statistical regression regarding house sale prices within King County, WA. The independent variables concern such data as number of bathrooms and whether the property has a nice view and I used them to predict what price any given house in the data actually sold for, using a support vector machine (or SVM) library. Of course there are very few perfect algorithms and here there is no exception. So how good did I do? Well, here's likely the best outcome on machine learning hub Kaggle, with over 700 up-boats:

Likely the best outcome on the data

His best model had a RMSE of 150177.258 US dollars. If you run my code you should get a RMSE of 166832.8 US dollars. I may not be the absolute best in the arena of machine learning but scoring only ~11% more than the true pros on the test set—remember, it's like golf: the lower, the better—isn't too awful for jack of all trades like me.