旅游人数的科学准确预测对旅游管理部门制定相关政策有着重要 的参考意义, 寻求科学的预测模型是保障预测结果准确可靠的关键. 文章以中 国国内旅游人数为研究对象, 首先,采用奇异谱分析对原始序列剔除噪声; 其次, 基于新序列利用3种单一模型对旅游人数进行建模预测分析; 最后, 构建了两种集成模型, 对各单一模型的预测结果进行综合分析. 研究发现, 文章所构建的SSA集成模型的预测精度极高且预测表现较稳定, 两种集成法在测试集上相对误差分别仅为0.26\%和0.34\%. 文章研究证明了奇异谱分析和集成预测模型在旅游预测中的价值, 对于科学预测地区旅游业发展具有指导意义.

The scientific and accurate prediction of the number of tourists has important reference significance for tourism management departments to formulate relevant policies. Searching for scientific prediction model is the key to ensure accurate and reliable prediction results. This paper takes the number of domestic tourists in China as the research object. First, it uses singular spectrum analysis to remove noise from the original sequence; second, it uses three single models to model and predict the number of tourists based on the new sequence; and finally, two integration models are constructed to comprehensively analyze the prediction results of each single model. The study found that the SSA integration model constructed in this paper has extremely high prediction accuracy and stable prediction performance. The relative errors of the two integration methods on the test set are only 0.26\% and 0.34\%, respectively. This paper proves the value of singular spectrum analysis and integrated prediction models in tourism prediction, which has guiding significance for the scientifically prediction of regional tourism development.