General Bayesian Network in Performing Micro-reality Mining with Mobile Phone Usage Data for Device Personalization

Personalization is an emerging issue in the digital age, where users have to deal with many kinds of digital devices and techniques. Moreover, the complexities of digital devices and their functions tend to increase rapidly, requiring careful attention to the questions of how to increase user satisfaction and develop more innovative digital products and services. To this end, we propose a new concept of micro-reality mining in which users’ micro behaviors, revealed through their daily usage of digital devices and technologies, are scrutinized before key findings from the mining are embedded into new products and services. This paper proposes micro-reality mining for device personalization and examines the possibility of adopting a GBN (general Bayesian network) as a means of determining users’ useful behavior patterns when using cell phones. Through comparative experiments with other mining techniques such as SVM (support vector machine), DT (decision tree), NN (neural network), and other BN (Bayesian network) methods, we found that the GBN has great potential for performing micro-reality mining and revealing significant findings.