Enhance Recommendation System for Hotel Groups

The hospitality industry has witnessed a notable increase in the utilisation of online booking platforms and the reliance on online reviews in recent years. This phenomenon presents a challenge for customers in their search for a suitable hotel that aligns with their specific requirements. Machine learning techniques were employed to develop a hotel recommendation system as a means of addressing this issue. The objective of this report is to elucidate the development process and evaluate the efficacy of the aforementioned system. The CRISP-DM approach was employed in conducting the present investigation. The system underwent training using a dataset acquired from kaggle, consisting of hotel data that had been organised and transformed into a format suitable for utilisation by machine learning models. The system employs a database including hotel attributes and customer evaluations in order to compute the several machine learning techniques. Several machine learning algorithms(Logistic Regression, K-Nearest Neighbors (KNN), XGBoost, LightGBM, CatBoost with categorical feature handling), resampling techniques (Random resampling, Equal class distribution resampling) and feature selection techniques(Select Best, Chi-square test) were employed of which XGBoost was chosen as the best algoritm. The study's findings indicate that the implementation of a machine learning-based hotel recommendation system has the potential to provide customers with valuable suggestions, hence enhancing their hotel booking experience. This study holds significance as it contributes to the field of hospitality by providing a practical resolution to the issue of hotel suggestion and addressing the existing knowledge gap about the utilisation of machine learning for enhancing hotel recommendations.