PBRTQC data cleaning method and system

Abstract

The invention belongs to the field of clinical data processing, and particularly discloses a PBRTQC data cleaning method and system. The method comprises data extraction, data preprocessing, data cleaning, data output and data security processing, the data cleaning comprises data interval truncation, and the data interval truncation is reference data interval truncation of healthy people or reference data interval truncation of personalized detection items; data filtering: judging whether the type of the truncated data is continuous data or discrete data, identifying and processing an abnormal value for the continuous data by adopting a Z-Score or IQR rule, and identifying and processing an abnormal value for the discrete data by adopting an occurrence frequency; deskewing treatment is carried out; and dealing with residual skewness and checking a transformation effect. According to the data cleaning method, the accuracy of clinical data can be improved, abnormity can be found earlier, unnecessary medical risks are reduced, and a reliable basis is provided for clinical diagnosis and treatment.

Images