Just a short doubt… in question 1 if we consider the buyer’s perspective will it not be a supervised, classification problem?
Merci.
If the buyer has a fixed budget and the problem is to predict if a house price is either above or below that given budget, it would indeed be a classification problem.
If the goal is to find houses that are cheap compared to their market value, one needs to predict prices as a continuous value, then aim for the real value to be below the predicted value.
Supervised learning means that:
- you want to predict the value of a specific yet unknown target variable (at prediction time) using the values from the variables you can always observe;
- you can collect training data both the input variables (“features”) and the target variable for other observations (often made earlier in the past relative to the prediction you want to attempt) of the same phenomenon (here the transactions of other housing on the same market).
If both conditions are met, then you can cast this task as a supervised machine learning problem.
The fact that your are a potential buyer, a seller, a real estate agency, employee of a market regulator or law enforcement agency, a data-journalist or an academic researcher does not really matter.