Knowledge discovery for interesting places for tourists in Johor Bahru, Malaysia

Abstract

Nowadays, Tourists are presented with a lot of online recommendation options before traveling. They often get confused in choosing specific places to travel and this is a time consuming process among all tourists across the globe. In this project, we crawled tourist profiles and interesting places in Johor Bahru from www.tripadvisor.com to discover clusters of customers with a different profiles, customers behavior, important feedback by tourists and useful knowledge in order to recommend appropriate places to tourists. This research includes two steps; in the first step, we clustered and applied ARM technique to uncover important knowledge about tourists and interesting places by Weka machine learning software. In clustering part we applied EM and K-Means algorithm and in association rules mining we used Apriori algorithm to find the rules between items in dataset. In the second step, we coded tourist’s comments, which are about interesting places in Johor bahru through Nvivo software. Results showed that, tourists could be clustered according to their preferences for instance, local people are not satisfied with the price of food in Legoland moreover, they prefer to travel with spouse and family with young children but foreigners like to travel with friends or business colleagues. Also, Legoland is one of the fix options for all male tourists aged between 25 to 34. Furthermore, Nvivo outputs shows that, Legoland has some affirmative and negative points. Local tourists believed that, assets of Legoland outweigh liabilities but some foreigners such as; Chinese and New Zealanders considered negative points like foods price and long queues. We believe that our two steps of analysis are powerful and results can be useful for tourism industry regarding to attract great bulk of tourists.