Singapore-Apartement-Pricing-Predictive-Model

This is a group project from my supervised machine learning class and not solely owned by myself. This project focuses on testing various regression methodologies for the best predictive model of apartment resale price in Singapore, based on official Singapore government data spanning from 2017 to 2022. The model was built on important features like floor area, storey range, etc. A lot of data preprocessing was done before testing with two encoding methods, different regression algorithms and their hyperparameters. Ultimately, the combination of one-hot encoding and linear regression likely produced the best results because one-hot encoding allowed categorical variables to be represented in a way that aligns well with linear regression’s mathematical assumptions, enabling the model to capture important variations in resale prices associated with each category. Linear regression, in turn, likely offered the right balance between model complexity and interpretability.