Sampling-based estimation method for parameter estimation in big data business era

Abstract

PurposeThis paper aims to present sample-based estimation methodologies to compute the confidence interval for the mean size of the content of material communicated on the digital social media platform in presence of volume, velocity and variety. Confidence interval acts as a tool of machine learning and managerial decision-making for coping up big data.Design/methodology/approachRandom sample-based sampling design methodology is adapted and mean square error is computed on the data set. Confidence intervals are calculated using the simulation over multiple data sets. The smallest length confidence interval is the selection approach for the most efficient in the scenario of big data.FindingsResultants of computations herein help to forecast the future need of web-space at data-centers for anticipation, efficient management, developing a machine learning algorithm for predicting better quality of service to users. Finding supports to develop control limits as an alert system for better use of resources (memory space) at data centers. Suggested methodologies are efficient enough for future prediction in big data setup.Practical implicationsIn IT sector, the startup with the establishment of data centers is the current trend of business. Findings herein may help to develop a forecasting system and alert system for optimal decision-making in the enhancement and share of the business.Originality/valueThe contribution is an original piece of thought, idea and analysis, deriving motivation from references appended.