In this paper, we consider click stream processing (most important use case of Big Data analytics). In future, predictive models and feature sets can be identified for other Big Data analytics workloads/data sets. More fine-grained predictions and approximation of the file’s computational profile can be generated.