Data Mining of International Tourists in Thailand by Two Step Clustering and Classification

Nowadays, the tourism industry is growing up rapidly. Market segmentation is an important tool to define marketing strategies which is driving to the business goals. This study proposed a data mining technique for tourist segmentation including (1) two step clustering and (2) classification of international travelers in Thailand. Two step clustering method, in the first step Self-Organizing Map (SOM) used for determining the appropriate group's number of tourists. The performance of K-Means was the best among three candidate clustering algorithms: K-Means, FCM and SOM. They were evaluated by using Silhouette, Root Mean Square Standard Deviation (RMSSTD) and R Square (RS). Then, K-Means used for partitioning tourist data in the second step. The statistical analysis in each segment was performed. The experimental results indicated that two step clustering method provided 9 natural segments. The second stage, the tourist classification by Multilayer Perceptron (MLP) indicated the highest precision as 97.43%. Tourism industry can use the result of this study for guiding to the marketing plans or promotion campaigns.