朴素贝叶斯在居民出行调查中的应用

Usage of Naive Bayes in Residential Trip Survey

为充分发挥居民出行调查结果的效用,采用贝叶斯分类法对天水市居民出行调查结果进行分类研究。主要在居民收入水平、是否拥有小汽车、出行时间等方面建立了小汽车的使用模型,模型精度达到94.1%。模型结果表明小汽车使用者受到居民收入水平、是否拥有小汽车的影响较大,同时小汽车使用者的忠诚型非常高,交通拥堵、出行时间拉长对小汽车的使用影响较小。下一步可以针对公交服务水平的改善、停车费用的增加等方面进行深入的模型研究。

In order to give full play to the effect of residenttrip survey, this paper uses Bayesian classification method to classify the results of Tianshui resident tripsurvey.In this paper, the model of caruseage is established in the aspects of residents＇ income level, car ownership and travel time. The model accuracy is 94.1%.The results show that car users are affected by residents＇ income level and car ownership, the car users＇ loyalty is very high, traffic congestion and travel time have less influence on car use. In the future, how the improvement of public transport serviceor increasing parking charges to affect car usage could be discussed.