论文部分内容阅读
本文基于百度指数进行旅游关键词的挖掘,运用决策树、bagging、随机森林和支持向量机四种算法模型对旅游关键词与平遥古城游客流量关系进行分析,并比较了各模型的拟合度、稳定性及预测效果。实证研究发现,随机森林模型稳定性最好,SVM模型拟合和预测效果最好,因此SVM模型可以作为最终的平遥古城客流量预测模型。
Based on the Baidu Index, this paper analyzes the relationship between tourism keywords and tourist traffic in the ancient city of Pingyao by using the four algorithm models: decision tree, bagging, stochastic forest and support vector machine, and compares the fitting degree, Stability and prediction effect. The empirical study shows that the stochastic forest model has the best stability and the SVM model has the best fitting and forecasting effect. Therefore, the SVM model can be used as the final passenger flow forecasting model in the ancient city of Pingyao.