Description
Abstract
According to the growth of tourists' demand for exploring authentic experiences in the natural environment and supporting conservation efforts, nature-based tourism is expanding sectors within the global tourism industry. National parks are the leading destination for naturalists who want to immerse themselves in the natural landscape's diversity, such as forests, coral reefs, and waterfalls. Knowing about the preferences of the tourists' activities within these destinations is crucial for effective tourism destination management and marketing strategies. This study aims to examine an association rule mining model in the field of machine learning approach to discover the set of activities most frequently made among tourists and preferences of the activities within the context of nature-based tourism in Phu Kradueng National Park (PKND) in Loei province, Thailand. The significance of this research is the potential to provide valuable information about the types of activities frequently chosen by visitors and take place at PKNP. Besides, it can aid marketers in planning and developing target marketing strategies to enhance the quality of tourist experiences.
This study constructed the data mining process based on knowledge discovery in databases (KDD) using an association mining model. This process is specific to data mining methods for pattern discovery and extraction, which includes. - (1) data understanding, (2) data preprocessing, (3) data modelling, and (4) data evaluation. This study focused on Phu Kradueng National Park (PKNP), Thailand. PKNP is in the Sri Than sub-district, Phu Kradueng district, Loei province, located in the northeast of Thailand. This park was designated the 2nd National Park in Thailand in 1962 and is presently administered by the Department of National Park, Wildlife and Plant Conservation (DNP). It covers an area of 348.12 kilometres and consists of a sandstone mountain with abundant flora and fauna and cliffs, grassland, streams, and waterfalls. The data were collected through a self-administered questionnaire survey via Google Forms with the respondents who had visited the PKNP. The questionnaire survey was developed based on the literature review and relevant past studies focused on tourist perceptions of activity components. After the questionnaire was developed, it was transformed into Google Forms to make it easier for the respondents to answer. Multiple responses were applied to the tourist activities performed at the PKNP. The respondents were recruited through social media platforms. It comprises Facebook and Twitter and is qualified based on the people who have travelled to the PKNP during the past five years since their last visit. The respondents were given a month to complete the questionnaire. The list of activities was provided by the national park management of PKNP, including activities such as hiking, camping, sightseeing, and volunteering. The attribute used is the activities that he or she participated in the PKNP with the value between 1 and 0, whereby the value 1 represents the tourist engaged with the activity, whereas 0 represents the tourist who did not engage with the activity. A total of four hundred and sixty-five (465) samples were analysed. Data preparation was performed after the questionnaires were collected. The data were recorded, transformed into a Microsoft Excel file, and mined using Rapid Miner software. The data was cleaned from missing values. There were 11 attributes in the dataset. The activity value in the dataset was recorded as binary data (0 or 1). The study employed the FP-Growth Algorithm within the association model to identify significant associations among tourist activities.
The results showed that sightseeing, camping, and picnicking were the most engaging activities at PKNP, with high support values indicating their popularity among visitors. These findings offer valuable insights into the preferences of tourists visiting PKNP, enabling tourism authorities and operators to tailor their offerings to meet visitor demands more effectively. The implications of this study contribute to the machine-learning approach to annotating PKNP's tourists' preferences of activities. By tackling artificial intelligence techniques such as association rule mining, researchers can uncover complex patterns within large datasets that may not be discernible through traditional manual analysis methods. It makes it straightforward to understand tourist behaviour and provides a foundation for tourism development. The findings of this study also have significant implications for tourism management and marketing strategies at PKNP. Besides, park managers can allocate resources more effectively by identifying the most engaging activities for visitors, developing tailored marketing campaigns, and enhancing the overall visitor experience.
Most importantly, the insights from this study can provide decision-making processes to the park manager in order to improve infrastructure development, conservation procedures and visitor management at PKNP. However, it is essential to acknowledge the limitations of this study. Further research is warranted to explore additional attributes such as length of stay and frequency of visit and to conduct seasonal trend analyses. Additionally, the study's findings may be limited to the specific context of PKNP and may not be generalisable to other nature-based tourism destinations. In conclusion, this study examines the effectiveness of machine learning approaches, specifically association rule mining, in unfolding patterns of tourist activities within nature-based tourism destinations like Phu Kradueng National Park. By providing valuable insights into tourist preferences and activities, this research contributes to developing more targeted and personalised tourism experiences, ultimately enhancing visitor satisfaction and supporting the sustainable management of natural resources.
