Study on the Method of Data Preprocessing for QAR Data

摘要

Data preprocessing is the key to proide quality data for QAR data warehouse. Usually there are lots of missing values, duplicate records and noise data in decoded QAR data, which seriously affect the efficiency of analysis of the QAR data. To deal with these problems, an architecture design of QAR data preprocessing based on subject is provided. The idea of design aims to provide a data analysis foundation for data analysis and mining, which covers the data quality improvement and data reorganization according to different analysis themes by data extraction, stuffing the missing attribute value and integration.