Digital file size computational procedure in multimedia big data using sampling methodology

The multimedia big data has tendency of fast growth over time span due to basic characteristics like volume, variety and velocity. Sample based estimates are used to compute the unknown population parameter. The multimedia big data is characterized various features who are prominent in terms of identification and analysis. Social media platforms are major sources of generating big data by virtue of communication among registered users in the form of text, video and images data. Registered users are also growing with drastic speed. When a user registers on a portal, default digital storage space is allotted by the system, who increases over time domain. A monitoring system is required to anticipate the increment and to alert managers of data centers for further enhancement of infrastructure. In the case of medical diagnostics, the CT-scan and MRI equipment produce the huge amount of scan files data while done over the large number of patients. These files are used to store in memory of the system for at least a prefixed duration. The digital file size of such reports, pixel densities and intensity of contents are the prime parameters of interest while comparing the quality of similar types of machines. Doctors and patients on social media platform used to exchange digital medical reports like X-ray, CT-scan, MRI, cancer diagnostics occupying the default storage. A guess value of digital file size can be helpful for the determination of expected amount of digital storage for users to be allocated to the medical processionals or other such. This paper presents sample based estimation methods for estimating the average file size over several time points. Confidence intervals are used as a tool of comparisons. A new simulation procedure is also suggested for comparative results of confidence intervals. At multiple optimum values of constant, the proposed sample based estimation methods perform better and the proposed simulation method is also result oriented. Strategy of using support information of other multimedia variable in estimation procedure is found useful and effective. Findings of the paper are numerically supported and significant percentage gain observed for proposed in the setup of multimedia big data floating over social media platforms.