Selection of Data Preprocessing Techniques and Its Emergence Towards Machine Learning Algorithms using HPI Dataset

To compute the frequent metamorphosis of the housing price, the House Price Index (HPI) is one of the effective indicators. Various methodologies are involved in the data processing of the current house prices, which are affected by multiple factors like house configuration, building class, air conditioning quality, etc. Remarkably, a greater number of research papers adopting classical machine learning approaches are introduced to estimate house sale prices accurately. Still, they barely regard the data processing techniques that make the data suitable for modelling more accurate house price forecasting models. This research contributes to a wide variety of adequate data preprocessing. It extensively highlights mechanisms like missingness of data, missing data handling, outliers, and feature scaling extensively to build efficient predictive models. Compre-hensive arguments have been broadly presented to portray the advantages and disadvantages of prevailed data preprocessing techniques at various distribution scenarios of variables in the house price data. The current research conclusions oblige the evolution of modern data-driven research in machine learning.