自然科学版学报

从初始数据源出发,总结了目前数据预处理的常规流程方法,提出应把源数据的获取作为数据预处理的一个步骤,并且创新性地把数据融合的方法引入到数据预处理的过程中,提出了数据的循环预处理模式,为提高数据质量提供了更好的分析方法,保证了预测结果的质量,为进一步研究挖掘提供了较好的参考模式.

Beginning from the initial data source,data preprocess routine technological process method is summed up,and at the same time,it is put forward that,gaining source data should be as a step of data preprocess.Datafusion is brought into data preprocess,and the data circulation preprocess pattern is proposed;it is a fairly good reference pattern for further studies in data mining,and provides much better analysis method to raise the data mass,and gave an important guarantee to forecasting the result mass.