K-means++ Algorithm Based on Spark Platform in Large-scale Tourism Data Mining in Cloud Computing Environment

With the acceleration of global economic integration and the rapid growth of tourism data, the processing and analysis of tourism data are facing severe challenges. In order to manage and analyze massive data effectively, it must use effective data mining technology. In order to solve this problem, this project plans to build a set of K-means++ algorithm for big data based on cloud computing and Apache Spark platform, aiming at improving the efficiency of big data clustering analysis. Spark is chosen as the development platform because it has strong storage capacity and scalability, and can effectively handle massive data. On this basis, a clustering method based on Kmeans is proposed, and the initial value is optimized, thus shortening the convergence speed and improving the stability and accuracy of clustering results. The implementation method includes first preprocessing large-scale tourism data, such as data cleaning and feature selection, then applying an improved K-means++ algorithm for data clustering, and finally analyzing the clustering results to identify potential trends and consumer behavior patterns in the tourism market. As the maximum number of iterations increases, the average intra cluster distance gradually decreases from 10.5 to 9.8, indicating that the clustering effect is gradually improving, that is, data points are more tightly concentrated within their respective clusters. This study provides an effective big data analysis tool for the tourism industry, helping it better understand customer needs, optimize service supply, and promote the development of personalized tourism products.