Enhancing House Price Prediction Accuracy using Deep Learning and Hyperparameter Tuning with AEDA

Traditionally, real estate agents and appraisers used their expertise and market knowledge to estimate house prices. However, due to the expansion of data sources and developments in data science, there was a growing interest in developing predictive models to automate and enhance the house price prediction process. This research introduces the Adaptive Exploratory Data Analysis (AEDA) tool, specifically the AEDA by frequency for categorical values, implemented in Python. The tool provides an API that maps usual AEDA tasks in statistical modeling for generating plots and statistics. It automatically generates insights and establishes standards and procedures. The frequency of users in various variables highlights the challenges in data processing pipelines for predicting house prices using Machine Learning (ML) and Deep Learning (DL) techniques. The chapter also discusses the evaluation of the hyperparameter tuning model for house pricing in the subsequent chapter. Model building, in the context of ML, involves creating a model that can learn from data without explicit instructions. Regression analysis is utilized to observe and resolve prediction issues. The suggested house price prediction model combines a joint self-attention mechanism with a specific feature parameter representation. While buyers and sellers focus on external features like building area, location, and construction year, the proposed model emphasizes internal features. It emphasizes the importance of hyperparameter tuning in ML and DL algorithms to improve model accuracy by minimizing residual errors during validation.