Sampling Technique for Content-Based Prediction in Big Data Environment

With the remarkable progress of big data analytics, people are paying attention to utilize a portion of it in prediction and forecasting. Although it is hard to recognize big data even then within the dimensions of volume, variety and velocity, one can understand the basic characteristics and identify them. A complex adaptive system refers to the ability of individuals in a system to communicate with other individuals as happens in the social media environment. People are using social media sites for the exchange of thoughts and expressions that ultimately generate big data at the cloud/ data centers. Social media is an adoptive system having a large number of interconnected users linked at a common platform. While communication, files of different formats are exchanged among users of different categories in terms of country, state, cast, age and financial levels. This paper considers the prediction about the average digital file size likely to exchange over time. The properties of the proposed method are discussed with the computation of 95% confidence intervals. A simulation procedure is used to directly compute the confidence intervals and found useful in catching the true unknown value.