[Paper Review] An Intelligent Data Analysis for Hotel Recommendation Systems using Machine Learning

Bushra Ramzan, Imran Sarwar Bajwa+2 more|arXiv (Cornell University)|Oct 15, 2019

Recommender Systems and Techniques42 references4 citations

TL;DR

This paper proposes a novel collaborative filtering-based hotel recommendation system that integrates sentiment analysis of user reviews and guest type profiling (e.g., solo, family, couple) to improve personalization. By combining lexical, syntactic, and semantic analysis for polarity detection and using fuzzy rules with Hadoop for big data processing, the system achieves higher precision, recall, and F-measure compared to traditional methods, with significantly improved response time.

ABSTRACT

This paper presents an intelligent approach to handle heterogeneous and large-sized data using machine learning to generate true recommendations for the future customers. The Collaborative Filtering (CF) approach is one of the most popular techniques of the RS to generate recommendations. We have proposed a novel CF recommendation approach in which opinion based sentiment analysis is used to achieve hotel feature matrix by polarity identification. Our approach combines lexical analysis, syntax analysis and semantic analysis to understand sentiment towards hotel features and the profiling of guest type (solo, family, couple etc). The proposed system recommends hotels based on the hotel features and guest type as additional information for personalized recommendation. The developed system not only has the ability to handle heterogeneous data using big data Hadoop platform but it also recommend hotel class based on guest type using fuzzy rules. Different experiments are performed over the real world dataset obtained from two hotel websites. Moreover, the values of precision and recall and F-measure have been calculated and results are discussed in terms of improved accuracy and response time, significantly better than the traditional approaches.

Motivation & Objective

• To address the limitations of traditional collaborative filtering in handling heterogeneous, large-scale hotel review data.
• To improve recommendation personalization by incorporating guest type (e.g., solo, family) as a contextual factor.
• To enhance accuracy and response time in hotel recommendation systems through intelligent data analysis using machine learning.
• To develop a scalable solution using the Hadoop platform for processing big data from real-world hotel websites.
• To integrate sentiment analysis of user reviews to extract meaningful hotel feature preferences.

Proposed method

• The system employs a hybrid approach combining collaborative filtering with opinion-based sentiment analysis to extract sentiment polarity toward specific hotel features.
• Lexical, syntactic, and semantic analysis techniques are used to identify sentiment polarity in user reviews for each hotel feature.
• Guest type profiling (solo, family, couple, etc.) is integrated as an additional input to tailor recommendations.
• Fuzzy logic rules are applied to map guest type and sentiment scores to hotel class recommendations.
• The system is implemented on the Hadoop big data platform to manage large-scale, heterogeneous datasets from two real hotel websites.
• Precision, recall, and F-measure are computed to evaluate performance against baseline collaborative filtering methods.

Experimental results

Research questions

• RQ1 How can sentiment analysis of user reviews improve the accuracy of hotel recommendations?
• RQ2 To what extent does incorporating guest type (e.g., solo, family) enhance personalization in hotel recommendation systems?
• RQ3 Can a hybrid approach combining collaborative filtering with sentiment analysis outperform traditional CF methods in terms of precision and response time?
• RQ4 How effective is the use of fuzzy rules in mapping guest profiles and sentiment scores to hotel class recommendations?
• RQ5 What performance gains are achieved when using Hadoop for scalable processing of heterogeneous hotel data?

Key findings

• The proposed system achieves significantly higher precision, recall, and F-measure compared to traditional collaborative filtering approaches.
• Response time is notably reduced due to efficient data processing using the Hadoop platform.
• Incorporating guest type as a contextual factor improves recommendation personalization and relevance.
• Sentiment analysis based on lexical, syntactic, and semantic techniques effectively captures user opinions on specific hotel features.
• The integration of fuzzy rules enables robust mapping of user profiles and sentiment scores to appropriate hotel classes.
• Empirical evaluation on real-world datasets confirms the system's superior performance in accuracy and scalability.