基于LSTM的游客预测  

Visitor Forecast Based on LSTM

作　　者：顾永鹏 覃迪波 章详 崔朝翔 周才英[1] 

机构地区：[1]江西理工大学理学院,江西 赣州

出　　处：《应用数学进展》2023年第5期2143-2151,共9页Advances in Applied Mathematics

摘　　要：随着GDP不断的增长,越来越多的人们在节假日愿意将时间用来旅游。景区游客的数量急剧增长和需求的集中,极易造成区域交通的拥堵或景区人挤人、游客之间产生安全事故等现象。严重影响了游客的游玩体验,同时不利于旅游消费和旅游业的可持续发展。为了改善这种现状,本文对各个景区的游客数量按照时间进行预测,通过数据建立模型给游客提供旅游建议。针对上述问题,本文先搜集各个景区旅游数据,在数据量足够大的前提下,使用深度学习中的LSTM模型对数据集进行训练,通过前向传播和反向传播训练模型,从而得到预测结果。然后按照LSTM预测的游客人数,应用在旅游服务平台上,以指导游客错峰出行,提升游客的游玩体验,并有效帮助景区管理人员对商品和人员进行安排。With the continuous growth of GDP, more and more people are willing to spend their time traveling during holidays. The number of tourists in scenic spots increases rapidly and the demand is con-centrated, which is easy to cause regional traffic congestion, crowded scenic spots, safety accidents between tourists and other phenomena. It has seriously affected the tourists’ play experience and is not conducive to the sustainable development of tourism consumption and tourism. In order to im-prove this situation, we forecast the number of tourists in each scenic spot according to the time, and build a model through the data to provide tourists with travel suggestions. To solve the appeal problem, we first collect tourism data of each scenic spot, and on the premise of sufficient data, we use the LSTM model in deep learning to train the data set. Through forward propagation and back propagation training model, we can get the predicted results. Then, according to the number of tourists predicted by LSTM, it is applied on the tourism service platform to guide tourists to travel off-peak, improve their playing experience, and effectively help the scenic spot management per-sonnel to arrange commodities and personnel.

关 键 词：LSTM模型 CNN模型 ARIMA模型 SPSS软件 

分 类 号：F59[经济管理—旅游管理]