Multi-level Frequent Location Sequences Mining in RFID Database

RFID technology has been applied in more and more fields, such as supply chain management, product tracking and so on. Now many research focus on how to manage and analyze RFID data set. Since location information, which contains multiple conceptual levels, is an important part of RFID data set, mining multi-level location sequences assists the decision-making. Unfortunately, little attention has been paid to mining multi-level sequential patterns. In this paper, we propose a bitmap-based algorithm that efficiently discovers multi-level frequent location sequential patterns. Our algorithm scan the raw data set once to build the bitmap of the bottom level and generate the bitmap of higher level from the current one of lower level, rather than excessively scanning the raw data set multiple times to build the bitmaps of every level. Experimental results have shown that our new algorithm maintains its performance advantage across a broad spectrum of test environments.