期刊名称：The International Arab Journal of Information Technology
印刷版ISSN：1683-3198
出版年度：2012
卷号：9
期号：3
出版社：Zarqa Private University
摘要：Nowadays, the vital parts of the business programs are the Data warehouses and the Data mining techniques. Especially these are vital in the RFID application which brings a revolution in business programs. Manufacturing, the logistics distribution and various stages of supply chains, retail store and quality management applications are involved in the RFID technology in business. A large volume of temporal and spatial data is generated by the ubiquitous computing and sensor networks of RFID and these are often generated with noises and duplicates. The noises and duplicates in the RFID data declare the need of an effective data warehousing system. The warehousing system has the responsibility to provide proper data cleaning technique to clean the dirty data which occurs in the applications. Also, the cleaned data has to be transformed and to be loaded properly so that they can be stored in the database with minimum space requirements. In this paper, we propose a novel data cleaning, transformation and loading technique which makes the data warehousing system employed for any RFID applications more effective. The chosen RFID application is tracking of goods in warehouses using RFID tags and readers, one of the significant RFID applications. The data cleaning is performed based on the probability of each RFID tag’s response and the window size which is made adaptive. The window size changes on the basis of the occurrence of the dirty data and hence the cleaning is more effective. The purified data is transformed in a special structure in such a way that the ware house can have only the tag IDs which are under transaction and the time of interrogation in the size of bits. The transformed data are loaded into the warehouse using the proposed loading technique in a dedicated tabular format.
关键词：Data warehousing system; data cleaning; data transformation; data loading; dirty data; radio frequency Identification (RFID); and suspicious tags
