Real-estate market will be part of every one of us concern in some phase of our life, in the era of technological progress questions comes up about the role of the brokers in the market and how could we know that we made the best deal for us.
Our project tries to predict real estate prices for future deals, the prediction based on analyzing a data-base that contains details about deals made in the past.
We used diverse approaches and tools of machine learning, for every method we used we tried to optimize the result and bring the prediction error to the minimum.
We started with linear and logistic approaches and start to combine multiple algorithms and build a mashup that suit best for our purpose.
In addition, we tried to use deep learning approach and built neural network classifier.
Finally, we compare the results and improve the classifiers based on our conclusions.
Supervisor Jonatan Chernyak