Real-estate-in-Bangalore

Classification and Prediction problem

Data

The data consists of features of real estate in different areas of Bangalore. It was pre-processed for convenience. The original data can be found here.

Variables:

• availability: is the property available immediately (1) or in the near future (0).

• total_sqft: the area of the property in square feet (1 foot = 30.54 cm).

• bedrooms: the number of bedrooms in the property.

• bath: the number of bathrooms in the property.

• balcony: the number of balconies in the property.

• rank: the ranking of the neighborhood in terms of average price (1 is the highest).

• area_type: is the property type a built up area (B) or plot area (P).

• price in rupees: the price of the property.

Split:

• Train: rows 1-8040.

• Validation: rows 8041-10050.

• Test: rows 10051-12563.

Section A

1. Decision Tree: Implement a Decision Tree (classifier and regressor) algorithm in Python.
2. AdaBoost: Implement an AdaBoost (classifier) algorithm in Python.

Section B

1. Classification: Use both models from section A and predict the area type (B, P), using all the features in the dataset.
2. Regression: Use the decision tree model from section A and predict the price of a property, using all the features in the dataset.

Section C

1. Sklearn Models: Implement the models (including hyperparameter tuning) from section B using built-in function from Sklearn.
2. Comparison: Compare the result of your program and the built-in Sklearn models in terms of metrics and runtime. If there are differences, suggest an explanation.

Section D

1. Gradient Boost Regressor:
   • Implement a Gradient Boost Regressor algorithm in Python.

   • Run the gradient boost algorithm on the given data to predict the price of a property.

   • Compare the algorithm’s performance to the built-in Sklearn model and the previous models that you implemented.

2. Classification Metrics:
   • Report the sensitivity and specificity metrics of section B(1).

   • Is there a significant difference between the scores? Suggest an explanation to why that may be the case.

   • Suggest and apply a method to improve the scores.

3. Performance:
   • Additional bonus points (up to 5) will be given for outperforming other students (in terms of metrics). Make sure to provide an explanation.