Enhancing House Price Prediction and Accuracy using Particle Swarm Optimization

Predictive models for determining the sale price of houses in cities like Bengaluru is still remaining as more challenging and tricky task. The sale price of properties in cities like Bengaluru depends on a number of interdependent factors. Key factors that might affect the price include area of the property, location of the property and its amenities. In this research work, an analytical study has been carried out by considering the data set that remains open to the public by illustrating the available housing properties in machine hackathon platform. The data set has nine features. In this study, an attempt has been made to construct a predictive model for evaluating the price based on the factors that affect the price. Modeling explorations apply some regression techniques such as multiple linear regression (Least Squares), Lasso and Ridge regression models, support vector regression, and boosting algorithms such as Extreme Gradient Boost Regression (XG Boost). Such models are used to build a predictive model, and to pick the best performing model by performing a comparative analysis on the predictive errors obtained between these models. Here, the attempt is to construct a predictive model for evaluating the price based on factors that affects the price.