摘要： 多年积累的实测糖尿病数据是一项重要的不可再生资源,对其数据的有效管理、处理、利用是一项非常必要和有意义的工作。该文就经过长期积累的2万多人次的调查和实测数据进行了有效的前期处理,其中涉及健康调查表内容以及原始数据的正确录入,对存入数据库的原始数据实施的数据清理、数据变换和数据规约3项预处理技术。2型糖尿病数据的预处理可以消除数据中的噪声、不完整和不一致性,实现数据的归一化和有效压缩,从而形成高质量的数据,可使数据的再处理(统计、数学建模、数据挖掘等)更加有效。此外,经预处理过的数据还可以作为一种重要的资源加以长期保存。

关键词: 2型糖尿病, 数据清理, 数据规约, 数据预处理

Abstract: The practice data of the type 2 diabetes that accumulated by many years is not re-genesis, so the effective management and use are an important and meaningful work. The pre-processing technologies of more then twenty thousands peoples surveying and measuring type 2 diabetes data are discussed, which involve the recording of the data, the sorting out of the data and the translation of the data and the stipulations of an agreement of the data. The noise, the non-integrity and the non-consistency of the data can be disposed by the pre-processing technologies, at the same time, the data can be classified and compressed, so the high quality data can be obtained. Beside these, the high quality data is and important natural resources to be conserved timeless.

中图分类号: 

• TP319