A Survey on Dynamic Structure Embedded Online Multiple-Output Regression for Streaming Data
International Journal for Research in Applied Science and Engineering Technology
This paper introduces a new strategy called Multiple Output Regression for Streaming data named as MORES. It works on streaming information and regression coefficient. The goals of MORES consist of four important perspectives: 1) It uses the mahalanobis matrix to find dissimilarity between the previous regression coefficient and the current regression coefficient to update the model. 2) It uses Mahalanobis distance for finding prediction mistake and increasingly finds remaining errors, it
... t use Euclidian distance. 3) It separates the very important data for better prediction and for estimating errors. All vital data stored in the memory and MORES use these data for event purposes. Labeling of information is used in the multiple-output regression for developing information. Mainly lebeling is used to reweight data for finding errors. 4) Different strategies are introduced in the MORES to develop the system architecture. Eigenvalue decomposition method is used to measure elements of the data sources.Eigenvalue decomposition method is introduced for the situation when the size of the data becomes large or for large streaming information. In this way online multiple-output regression uses the techniques of machine learning system for modeling correlated data stream and predicting multidimensional related data stream and it always provides model refinement.