Picking A Model To Predict Future House Prices in the US

Ladies and gentlemen, I have been well aware on the alleged great migration within the United States of America.

I've witnessed the Northeastern exodus to Florida.

I've seen videos of the California exoduses to Texas, Nevada, and Arizona.

The main reasons being are increased state taxes, increased costs of living, as well as state politics from their old places of living to evade.

Now, in recent years, the cost of living and other fees have risen everywhere, especially in the hottest destinations to move to:

• Texas: house prices have risen dramatically due to high demand and low supply and labor; mainly in the Austin, Dallas-Fort Worth, & Houston metroplexes.
• Georgia: house prices have risen in the Atlanta metroplex
• Florida: house prices have risen statewide, especially when you factor in the global warming conditions there.

Around a month ago, I took the time to do some digging on this ordeal using data science and machine learning to pick which model is best suited for predicting future house prices in the country. Here's what I have found so far:

If you take a look at the graph to the left, you'll notice the value of American homes skyrocket over the past decade.

If you take a look at this graph to the left, you'll notice the value of adjacent American homes also skyrocket over the past decade. Please take note that the adjacent housing price is the price of a neighboring home right next to the one that you're targeting.

I've tested ten machine learning models to see which one is best suitable for prediction.

For the first six models, I've utilized the following:

• Logistic Regression
• LASSO
• Elastic Net
• K-Nearest Neighbors
• Classification & Regression Trees
• Supporting Vector Machines

Now let's scale them.

All of the ten models are measured by a negative mean square error (MSE). The model with the least negative MSE will be the one that we'll use for a prediction. Now let's test out the remaining four ensemble models:

• AdaBoost
• Gradient Boosting Method
• Random Forest
• Extra Trees

Based on all of the data shown, the Extra Trees model will be used as a predictor.

For more information on this dataset, please visit my GitHub page below.

#inflation #datascience #machinelearning #realestate