Restaurant Recommendations Through Natural Language Processing Based on User Rating

Abstract

This research paper presents a comprehensive study on the development and implementation of a novel Restaurant Recommendation System (RRS) leveraging machine learning techniques and geographical data. The system integrates user preferences, location information, and historical restaurant data to offer personalized recommendations. our research endeavours to contribute to the domain of personalized restaurant recommendation systems. We propose a comprehensive system that amalgamates machine learning techniques, geographical analysis, and a user-friendly graphical interface to offer tailored dining suggestions to users in urban settings. Our approach integrates data pre-processing techniques to handle duplicate entries and encodes categorical variables for enhanced model interpretability. The predictive model, a linear regression algorithm, strives to estimate restaurant prices based on features such as cuisine type, location, and dish category. Geocoding is employed to calculate distances between user-specified locations and recommend establishments within a user-defined radius. The system leverages a Random Forest Classifier to enhance the classification of dish categories, contributing to more precise recommendations. Our system not only predicts restaurant prices but also identifies the most popular establishments based on user-defined parameters. By combining predictive modeling, geographical analysis, and classification algorithms, we aim to create a robust restaurant recommendation system that aligns with the evolving expectations of modern consumers.