Targeted Advertisement Based On User Behavior For IPTV

The trend of customers' demand presents differentiated,personalized and diversified,through segmenting the market precisely,targeting and analyzing consumers' demand and behavioral characteristics,promoting target consumers their needed products and services can greatly improve advertising efficiency,reduce advertising costs and minimize the interference for the audiences.Traditionally,advertising has not paid any attention to the behavior characteristics of the users,which makes the delivery of advertisements lack of transparency and precision.Due to the development of artificial intelligent technology and the emergence of large-scale data,personalized advertising will certainly become the main trend of future advertising development.This paper presented a novel algorithm of targeted advertising based on user behavior with regard to how to distribute ads according to users' interest and preferences.The algorithm developed user behavior model based on the similarity between programs to estimate users' interest and preferences,and improved the efficiency of the traditional advertisement through considering both the users' heterogeneity and preference variability.Meanwhile,we considered the day of the week effect and holiday effect on users' view number,and got better results by the fixed effect and mixed effect model than traditional time series forecasting.