基于数据挖掘的汽车排放检测结果与使用特点关联性分析

Correlation Analysis of Vehicles Test Results and Usage Characters Based on Data Mining

本文利用SQLServer2008作为数据库管理及数据挖掘平台,并建立了一个基于约10万条在用汽车排放检测数据的决策树挖掘模型。数据挖掘结果指出,使用年限与车辆排放状态有着密切的关联性。随着使用年限的增加,在用汽车排放检测不合格率也逐步增加,排放检测的不合格率从使用年限区间(一)的2.96%逐渐增加至区间(五)的18.31%。另外,使用年限在6年以内,营运车的排放检测不合格率是非营运车的排放检测不合格率的2倍以上,因此应分别对这两种不同使用性质的车辆采用不同的检车周期,非营运车辆的检车周期应放宽于营运车辆,以减轻非营运车车主的检车负担以及节省相关费用;在使用年限大于6年,可以采用同一检车周期以便于车辆管理部门的管理工作。

This paper uses SQL Server 2008 as database management and data mining platform for establishing a decision tree mining model based on about 100 000 in-use vehicle emission detection data.The data mining models show that service life is closely related to vehicle emission status.With the increase of service life,the unqualified rate of emission detection of in-use vehicles is gradually increasing,the unqualified rate increasing from 2.96% in the interval(1)of service life to 18.31% in the interval(5)of service life.Besides,within 6 years of service life,the unqualified rate of emission detection of commercial vehicles was more twice higher than that of the non-commercial vehicles;therefore,should be used two different vehicle inspection cycles adopted for these two kinds of vehicles with different usage nature.The inspection cycle of non-commercial vehicles should be relaxed to reduce the inspection burden of non-operating vehicle owners and save related costs.With the service life is longer than 6 years;the same vehicle inspection cycle can be adapted for all kinds of vehicles to facilitate the management of vehicle management departments.