Young families can have a hard time navigating through the complexity and risks of real estate markets.
Is the price of a house in line with the market?
What is the amount that I should offer for a specific house?
Those planning to sell houses also need to engage one or more real estate brokers to estimate their house price; these, however, have an interest of their own in the process. Yet, when wanting to sell one’s house, one needs to know whether the house price estimate is in line with market prices.
Artificial intelligence can help to shed light on some of these questions. To demonstrate that, we used two machine learning algorithms to forecast house prices based on their specific characteristics. For that we used publicly-available web-scraped data from real-estate websites, which contain house offers in Bremen (where we are based).
Since the process behind the curtains can be quite complex, we packed everything into a web application (in short, ‘web-app’). This application, which we called “ImmoBot” simply requires users to specify the house characteristics and will then provide them with the estimates in no time. ImmoBot also presents the used dataset and plots, which we will be expanding from time to time.
Price forecasts are based on regression analysis from a gradient boosting algorithm and from a Random-Forest algorithm, which are common machine learning tools. Since the forecasts from each algorithm differ, we added the average of the forecast estimates as an additional piece of information.
Gradient boosting and Random Forest are decision tree-based ensemble models. In gradient boosting, a shallow and weak tree is first trained, and then the next tree is trained based on the errors of the first one. The process continues with a new tree being sequentially added to the ensemble and the new successive tree correcting the errors of the ensemble of preceding trees. In contrast, random forest is an ensemble of deep independent trees.
We hope that this helps as a simple example of how artificial intelligence can be applied while helping our Bremen users navigating the complex real estate markets.
Please share this web-app and let us know in case you have any suggestions or questions.
Disclaimer: Please notice that the price forecasts are merely illustrative. Neither movimentar GmbH nor any person acting on their behalf may be held responsible for the use that may be made of the information presented here.