Future HDB Resale Prices Predictive Model Based on Historical Values
dataset - https://www.kaggle.com/datasets/teyang/singapore-hdb-flat-resale-prices-19902020
This Notebook at ./information_extraction.ipynb extracts information from the dataset by answering the following questions: Questions:
- What is the location, room type of the top 5 town of resale flats transaction? Answer : Sengkang, Jurong West, Woodlands, Tampines, Yishun
- for resale flat with story range of 10-20 with number fo room >= 3, what is the location of the most affordable resale
- What is the highest and lowest resale price in each town with number of room ranges between 10 and 15
- what is the highest and lowest resale of every sales
- what is the median resale price for story range of 3-4
Objective: Create a predictive model of the resale price so that property agents are able to estimate the price range.
Inputs - the different features Output - a hdb resale price (continuous number) Model - Regression Loss function - BCE
- find correlation between features
- ensure balanced dataset
TODO
- change everything to numerical
- try one model ML
- see how to preprocess more
- try NN