An Efficient Mobile Data Mining Model

Mobile Data Mining involves the generation of interesting patterns out from datasets collected from mobile devices. Previous work are frequency pattern [3], group pattern [9] and parallel pattern [5]. As mobile applications usage increases, the volume of dataset increases dramatically leading to lag time for processing. This paper presents an efficient model that uses the principle to attack the problem early in the process. The proposed model performs minor data analysis and summary early before the source data arrives to the data mining machine. By the time the source data arrives to the data mining machine, it will be in the form of summary transactions, which reduces the amount of further processing required in order to perform data mining. Performance and evaluation shows that this proposed model is significantly more efficient than traditional model to perform mobile data mining.