Predicting house prize
=======================

Tell your own story, here is the example ....

You have a client who would like to buy a house in a given district
(zipcode) because he likes schools there and work not far from it.
He is looking for a house with are least 3 bedrooms and 2 bathrooms
and his budget is between XXk to YYk USA $. 
He likes the waterfront but can go without it. Prefers if the
sizes of nearby houses are quite similar.

--> perform exploratory analysis of the houses in the database
    sold in that zone
--> try out simple linear regression model based on three features
    selected with greedy algorithm (nested selection)
--> try out non-parametric regression based on at least 5 nearby houses,
    use boxcar kernel, you might try out also different kernels.

Present offer of 2-5 houses for consideration with your client. Discuss
if the prizes are reasonable according to your models or houses
are too priced or are a good buy.

Be creative but don't get overwhelmed with sophistication of the code.
Use your intuition and common sense in preparing offer for your client.