DATA MANAGEMENT ENVIRONMENT PROPERTIES INVESTIGATION AND TIME ESTIMATION OF LARGE DATA SET TRANSFER

The article considers the task of estimating time transmission large data sets through a distributed management environment based on self-created neural network model and investigating properties this using statistical analysis methods. For initial analysis, metadata for successful file transfers in system was obtained, variables that affect transfer were transformed highlighted. Different samples used to check whether results similar across available data. methods correlation regression are applied study environment. It found there is no clear between one input parameters. depends number external factors cannot be obtained metadata, but it possible partially investigate metadata. A two levels numerical categorical then combined into branch. RMSE metric value scatter plot display prediction compare target predicted values. performed calculations show satisfactory results;