Classification of Mixed Frequency Data Using a Novel Discriminative Dictionary Pair Learning Restricted by Ordinal Locality PROJECT TITLE : A Novel Discriminative Dictionary Pair Learning Constrained by Ordinal Locality for Mixed Frequency Data Classification ABSTRACT: One of the challenges that classification must overcome is the fact that the data are not gathered at the same frequency in all applications. We investigate the mixed frequency data in a new way and recognize them as a special style of multi-view data. In this particular form of multi-view data, the data for each view is collected at a different sampling frequency than the other views' data. This article presents a discriminative dictionary pair learning method for mixed frequency data classification that is constrained by ordinal locality (shorted by DPLOL-MF). This method combines the synthesis dictionary and the analysis dictionary into a single dictionary pair. This not only reduces the amount of computational work required due to the l0 or l1 -norm constraint, but it also allows for the sampling frequency inconsistency to be taken into account. The DPLOL-MF makes use of a synthesis dictionary to learn information about class-specific reconstruction, and it makes use of an analysis dictionary to generate coding coefficients by analyzing samples. Both dictionaries are utilized in the learning process. Specifically, the ordinal locality preserving term is used to constrain the atoms of dictionary pairs in order to further facilitate the learned dictionary pair's ability to be more discriminative. This is done in order to improve the accuracy of the learned dictionary pair. In addition to that, we devise a particular classification scheme in order to account for the varying sample sizes of the mixed frequency data. The purpose of this paper is to present a novel approach to the problem of classifying data with a mixed frequency distribution, and the experimental results demonstrate that the proposed method is effective. Did you like this research project? To get this research project Guidelines, Training and Code... Click Here facebook twitter google+ linkedin stumble pinterest Joint Hypergraph Embedding and Sparse Coding for Data Representation Unsupervised Domain Adaptation Using a Deep Ladder-Suppression Network