Analyzing User Behaviour of IPTV using Hidden Markov Model

Due to the innovative technologies in telecom industry and Internet Protocol TV (IPTV) became a reality. Many service providers across the globe are investing significantly to render Video on Demand (VoD) and other services through IPTV. Videos consume high bandwidth and they are to be multicast with acceptable service quality. Therefore it is indispensable for service providers to understand technicalities of IPTV and make necessary steps to have compatible infrastructure. Moreover IPTV user behaviour will have significant impact on the services rendered by the providers. Optimization of resources and services is possible provided the trends in the customer behaviour. As users of IPTV can have interactive usage of content delivered, their usage behaviour and events can affect quality of services. In this context, it is essential to characterize customer behaviour and analyze the same to have insights that can help service providers to have strategies to improve business by providing quality services to subscribers. In this paper our focus is on the usage behaviour of IPTV users besides their zapping patterns. We employed a stochastic model known as Hidden Markov Model (HMM) which has finite set of states and probability of state transmissions. Besides HMM has hidden states that capture user behaviour with respect to IPTV. The results of our research revealed the behaviour of users at different times and especially zapping rate. Index Terms – IPTV, user behaviour analysis, hidden Markov model, channel zapping ——————————  ——————————