大数据下浙江省公路货运量预测算法的设计与实现

Design and Application of Estimating Measures for Highway Freight Volume in Zhejiang Province via Large Data

运用大数据技术对公路货运量进行科学预测,能帮助各地区进行智慧交通管理。从浙江省统计局官方网站采集十年的数据,经过清洗和预处理,利用有监督的机器学习算法中常用的岭回归算法、朴素贝叶斯算法和KNN算法获得误差曲线,比较三种算法在相同测试集数据条件下获得的准确度和误差结果,寻找较准确的公路货运量预测手段。结果表明,岭回归算法对浙江省公路货运量的预测准确度较高。

Scientific estimate of highway freight volume via large data contributes to regional intelligent transport management. Data over a decade was collected from the official website of Zhejiang Statistic Bureau. Ridge regression algorithm, Naive Bayes Algorithm and KNN algorithm among supervised machine learning algorithm were applied to obtain the error curve after cleaning and preprocessing. Accuracy and error under the same test data condition were compared. Result shows that ridge regression algorithm has greater accuracy for estimating highway freight volume of Zhejiang Province.