关键词: 人工神经网络, 客源市场分区, 武夷山, 因子分析

Abstract: The clustering analysis of tourist market plays a lead role in the exploitation of touring market. Until now, although the quantitative methods are usually thought to be more dependable, more qualitative methods have been used to the clustering of tourist market due to the complexity of the behaving system of tourists. Artificial neural networks, simulating some characteristics of human brain, are proved to be effective models to the analysis of complex system. They are composed of many simple intercommunicating neurons that work in parallel to solve a special problem. They are much faster and effective than most conventional methods, especially for complicated behavior system that are impacted by all kinds of different factors, for once a network has been set up, it can learn in a self-organizing way which has much common in the mimic simple biological nervous system. In this paper, a soundly trained self-organizing feature map (SOFM) is employed to the clustering of tourist market of the Wuyishan, a famous scenic spot in the Fujian province, the southeast of China. First, concerning the available data, 11 variables that have influence on the demand of tourists are selected. Second, through the factor analysis, those 11 variables are compressed to 3 orthogonal factors. Third, useing these 3 factors as input variables, a SOFM is set up. When the neural network is trained appropriately, it classifies the data sets. The results show that 31 provinces or autonomous regions are classified into 3 groups, which is consistent with the results of the questionnaire survey in Wuyishan scenic spot. The results also indicate that SOFM is an alternative effective quantitative approach in clustering of tourist market.

Key words: artificial neural network, clustering of tourist market, Wuyishan, factor analysis