An Efficient Pattern Mining Approach for Progressive Databases

Abstract

Although there have been many recent studies on the mining of sequential patterns in a static database and in a database with increasing data. Newly arriving patterns may not be identified as frequent sequential patterns when generating sequential patterns due to the existence of old data and sequences. Recent studies have focused on mining sequential patterns in databases but often overlook the impact of deleting old data. This oversight can lead to newly relevant patterns being missed and outdated ones persisting. To address these, we proposed a model for sequential pattern mining in progressive databases, where data can be static, inserted or deleted. This approach accommodates the dynamic nature of data by using sliding windows that continuously update, reflecting the period of interest. In fields like medicine, analyzing temporal symptom-disease patterns aids in diagnosis and preventive care. Similarly in web server data, identifying frequent user access patterns informs system design and marketing strategies. Our model accounts for real world scenarios, such as items support in data streams, enhancing applications like market basket analysis. newline