基于卷积神经网络的人口流量预测

Prediction of Population Flow Based on Convolutional Neural Network

预测交通流量对于交通管理和公共安全具有十分重要的意义,准确的预测可以预防交通拥堵、防止踩踏事件的发生。同时,交通流受到各种复杂的因素的影响,例如多变的天气、休息日与节日,所以预测人口流量有一定的难度。文章采用了卷积神经网络对整个城市的人口流动进行建模,并训练模型使之可以用来预测一个城市人口的流入以及流出。使用卷积神经网络可以很好地抓取交通流的特性与不同区域的人口流入流出对周边区域的影响,从而准确预测下一个时刻整个城市各个区域的人口数量。最后,可以该卷积神经网络得到的结果进一步结合外部的复杂因素,来预测每个地区最终的人群流量。在这里使用了著名的北京交通出租车数据集对方法进行了验证,证明方法有不错的效果。

Forecasting traffic flow is very important for traffic management and public safety.Accurate forecasting can prevent traffic congestion and trample events.At the same time,traffic flow is affected by various complex factors,such as changeable weather,rest days and festivals,so it's difficult to predict the population flow.In this paper,the convolution neural network is used to model the population flow of the whole city,and the training model can be used to predict the inflow and outflow of a city's population.Using convolution neural network can grasp the characteristics of traffic flow and the impact of population inflow and outflow in different regions on the surrounding areas,so as to accurately predict the number of population in each region of the city at the next moment.Finally,the results of the convolution neural network can be further combined with external complex factors to predict the final population flow in each region.The method is validated by using the famous Beijing traffic taxi data set,which proves that the method has a good effect.