A close friend living overseas earned tens of millions of yen by investing in real estate. I want to spread my business, so I invite you to become a business partner. He spends money and creates a pattern (model) for you with data analysis skills to predict the price of a property.
You asked, "How do you predict and judge the price of a property?" I answered, "It's intuition." Isn't it possible to create a price forecast model intuitively by machine learning? Hearing in more and more detail, you somehow found out.
Property prices are divided into two groups.
In other words, the property price is predicted in the following two steps.
This procedure is called a fitting model or training model. The data that fits the model is called training data.
Use the fitted model to predict the price of the property you are investing in.
Which decision tree fits the training data?
If you think about it normally, decision tree 1 seems to be better. Larger properties are more expensive than smaller ones, right? However, the property price is not determined only by the size. There are many other factors that affect property prices (location, year of construction, etc.). The decision tree including other factors is as follows.
Choosing the path that suits the characteristics of the property, the estimated price is at the bottom node of the decision tree. The bottom node with the predicted price is called leaf.
The number of branches and leaf value of the decision tree are determined by the data. Next time, let's analyze the data.
Other related articles are here.
Recommended Posts