基于BP神经网络的自行车道健康影响评价

Health Impact Assessment of Cycleway Based on BP Neural Networks

慢行交通可以增加出行者的身体活动，但也会增加空气污染以及交通事故的暴露量。在现有的“健康影响路径（health pathway）-暴露反应关系-健康影响终点（health endpoint）”方法基础上，运用BP神经网络建立了健康影响评价计算模型。并以英国利兹市的5条自行车道项目数据对BP神经网络进行训练，训练结果显示误差大约在1％。运用劫I练后的神经网络模型对新建成的利兹-布拉德福德自行车快速路进行评价结果表明：从2016至2020年，每千人伤残生命调整年（DALYs）减少2．9996天至6．1561天，每千人全因死亡数减少0．09至0．14人，而健康收益主要来自慢行交通带来的运动量增加。

Active travel can increase physical activities but can also increase the exposure to traffic incidents and air pollution. Building health impact assessment model （HIA） with the ＂health pathway to exposure-response relationship to health endpoint＂ process and BP Neural Networks. 5 existing cycle paths were used to train the BP Neural Networks, the training results show that the error were about l%,which was acceptable. Conduct health impact assessment of the new built Leeds-Bradford Cycle Superhighway. The results show that from 2016 to 2012, DALYs per thousand people fall from 2.9996 days to 6.1561 days, all cause mortality decrease from 0.09 to 0.14, and the main contributor of health benefit is physical activities increasing from active travel.