THE MULTIPLE TIME SERIES CLINICAL DATA PROCESSING USING OPTIMIZATION ALGORITHM

Abstract

Author(s): Priyanga M1, Dr. K. Sasi Kala Rani2M.E.,Phd., Yamuna Devi S3, Pavithra M4

The main objective of this research is to discover patient acuity or severity of illness has immediate practical use for clinicians.The use of multivariate time series modelling along with multiple model is evaluated. As large-scale multivariate time series data become increasingly common in application domains, such as health care and traffic analysis, researchers are challenged to build efficient tools to analyze it and provide useful insights. In many situations, analyzing a time-series in isolation is reasonable. Initially MMSVM is applied for multiple measurements of time series dataset to classify the accurate results. But it has problem along with imbalanced dataset and it leads misclassification results. In the existing system, to overcome the unbalanced data and classification performance used improved Particle swarm optimization algorithm (IPSO). The unbalanced dataset is handled by using the improved PSO algorithm and it reduced the irrelevant feature in the given time series data. In the proposed system, introduced Modified Artificial Bee Colony Algorithm (MABCA) to solve the multiple time series problems by increasing the selection of optimal feature information. The MABA algorithm is used to improve the most appropriate features globally and global optimization is increased as well as dimensionality reduction. The MABCA reaches lighter designs along with a better convergence rate. In the proposed system, the optimal features are increased by tuning the parameters. The proposed MABCA with Transductive SVM (TSVM) is used to improve the classification performance and Artificial Neural Network (ANN) is used for making long term multi-step prediction. It reduces the training process time and minimize the error rate more significantly. From the experimental result, the conclusion decides that the proposed system is superior to existing system. Application/Improvements: The findings of this work prove that the graph search based method provides better result than other approaches