Asian Journal of Engineering and Technology Innovation

Event Characterization and Prediction Based on Temporal Patterns Using Mrps in Dynamic Data System

Deepak Dharra, Abhishek Iyer, Nitika Gulhane, Sultana Parveen Khan, Vishal Dond

Computer Engineering Department, Indira College of Engineering and Management Parandwadi, Pune.

Abstract

Multivariate Reconstructed Phase space is a new method proposed for identifying and characterizing temporal patterns. Among the several existing methods this method is the most efficient and beneficial. This method along with Gaussian Mixture Model characterizes events in Dynamic data system. MRPS is an extension of original univariate phase space and can be applied in numerous research fields. As MRPS considers Multiple data sequences it has wide application in medical as well as in financial time systems. Nowadays , Pattern Identification of ECG is gaining importance since the classification algorithm in MRPS results accuracy in diagnostics. MRPS-GMM approach can also be applied in order to estimate the GDP(Gross Domestic Product ) of a nation’s economy.

Keywords:

Multivariate Reconstructed Phase space, temporal Patterns, Gaussian Mixture Model