论文部分内容阅读
通过分析原始数据和“千车故障数”计算的不合理性,提出了改进的千车故障数的计算方法,并对原始数据进行变换得到了合理的数据表.通过假设检验得出瞬时故障率是服从Weibull分布,在此基础上提出了单批次预测的概率模型.进一步,利用数据横向和纵向的相关性,在单批次概率模型的基础上,提出了基于样条拟合的多批次双向联合预测模型,这个模型保证了横向和纵向的有机结合,通过拟合汽车的质量参数,来进行联合预测,维护了数据的统一性.
By analyzing the original data and the unreasonable calculation of “the number of failure of a thousand cars”, an improved calculation method of the number of the failure of a thousand cars is put forward and a reasonable data table is obtained by transforming the original data.The instantaneous failure rate Is based on the Weibull distribution, and on this basis, a single-batch probability model is proposed.Furthermore, based on the single-batch probability model, the multiple-batch model based on spline fitting Time bidirectional joint prediction model. This model ensures the combination of horizontal and vertical. By fitting the quality parameters of the automobile, the model can predict the joint and maintain the data consistency.