论文部分内容阅读
准确的日客流量预测对旅游景区至关重要,但受各种因素影响,日客流量呈现复杂、非线性特点,文章提出了一种基于自适应遗传算法(adaptive genetic algorithm,AGA)的支持向量回归(support vectorregression,SVR)模型,利用SVR处理非线性预测的能力和AGA参数寻优的特点,实现旅游景区日客流量预测。最后以某旅游景区2008.3-2012.6最新日客流量等数据集为例验证AGA-SVR模型的预测能力,并与GA-SVR和BPNN的预测结果进行对比分析。实验结果表明:同GA-SVR、BPNN相比,AGA-SVR能够有效的实现日客流量预测,准确性更高,误差更小,同时也说明利用AGA进行SVR参数选择是有效可行的。
Accurate daily traffic forecast is very important to tourist attractions. However, due to various factors, the daily traffic is complicated and nonlinear. A new support vector based on adaptive genetic algorithm (AGA) Support vector regression (SVR) model, the ability of SVR to deal with nonlinear prediction and the optimization of AGA parameters are used to predict daily passenger flow of tourist attractions. Finally, the forecasting ability of the AGA-SVR model is verified by using the dataset of the latest daily traffic of a tourist attraction from January 2008 to June 2012 as an example, and compared with the predicted results of GA-SVR and BPNN. The experimental results show that compared with GA-SVR and BPNN, AGA-SVR can effectively predict daily passenger flow with higher accuracy and less error. It also shows that the AGA-SVR parameter selection is effective and feasible.