An Effective Hotel Recommendation System through Processing Heterogeneous Data

Abstract

Recommendation systems have recently gained a lot of popularity in various industries such as entertainment and tourism. They can act as filters of information by providing relevant suggestions to the users through processing heterogeneous data from different networks. Many travelers and tourists routinely rely on textual reviews, numerical ratings, and points of interest to select hotels in cities worldwide. To attract more customers, online hotel booking systems typically rank their hotels based on the recommendations from their customers. In this paper, we present a framework that can rank hotels by analyzing hotels’ customer reviews and nearby amenities. In addition, a framework is presented that combines the scores generated from user reviews and surrounding facilities. We perform experiments using datasets from online hotel booking platforms such as TripAdvisor and Booking to evaluate the effectiveness and applicability of the proposed framework. We first store the keywords extracted from reviews and assign weights to each considered unigram and bigram keywords and, then, we give a numerical score to each considered keyword. Finally, our proposed system aggregates the scores generated from the reviews and surrounding environments from different categories of the facilities. Experimental results confirm the effectiveness of the proposed recommendation framework.

A Hotel Recommender System Based on Multi-Criteria Collaborative Filtering

“…The studies can be classified into two classes: The first class includes the hotel recommendation systems that utilize numerical rating information. The second class comprises the hotel recommendation systems that utilize text review information [1,11,12,14,18,20,[33][34][35][36].…”

“…Experiments using a dataset from iranhotel.com show that the proposed method improves CF performance. Forhad et al [14] introduced a hotel recommendation framework that analyzes customer reviews and local hotel amenities to make recommendations. First, the system calculates scores based on the reviews of the hotel booking datasets.…”

“…To evaluate the proposed model, the TripAdvisor MC dataset [17] is used for the experimental validation. The dataset includes 28,829 multi-, a x P , (14) where U x,y is the set of travelers who have commonly rated hotels x and y. ,…”

“…However, travelers commonly find it difficult to obtain valuable information from a plethora of online evaluations, making decision-making even more complex. As a result, an effective recommendation system that acts as a decision-making system can be utilized in order to make effective use of online rating information on specific hotel features for decision-making [10,14,35].…”

“…Abbasi et al [1] coupled natural language processing and a supervised classification approach to assess sentiments and extract implicit features from several hotel reviews. Forhad et al [14] introduced a hotel recommendation framework that analyzes customer reviews and local hotel amenities to make recommendations. Nevertheless, hotel reviews are not always available, and processing reviews is a demanding natural language processing activity that requires time and effort from all involved parties [19].…”