arXiv subjects

Ramasubramanian Sundararajan

Publications and source records attributed to Ramasubramanian Sundararajan.

4 recordsLinked to original sources

arXiv2015

Grouping Entities in a Fleet by Community Detection in Network of Regression Models

This paper deals with grouping of entities in a fleet based on their behavior. The behavior of each entity is characterized by its historical dataset, which comprises a dependent variable, typically a performance measure, and multiple independent variables, typically operating conditions. A regression model built using this dataset is used as a proxy for the behavior of an entity. The validation error of the model of one unit with respect to the dataset of another unit is used as a measure of the difference in behavior between two units. Grouping entities based on their behavior is posed as a graph clustering problem with nodes representing regression models and edge weights given by the validation errors. Specifically, we find communities in this graph, having dense edge connections within and sparse connections outside. A way to assess the goodness of grouping and finding the optimum number of divisions is proposed. The algorithm and measures proposed are illustrated with application to synthetic data.

cs.SI↗

arXiv2014

Double Ramp Loss Based Reject Option Classifier

We consider the problem of learning reject option classifiers. The goodness of a reject option classifier is quantified using $0-d-1$ loss function wherein a loss $d \in (0,.5)$ is assigned for rejection. In this paper, we propose {\em double ramp loss} function which gives a continuous upper bound for $(0-d-1)$ loss. Our approach is based on minimizing regularized risk under the double ramp loss using {\em difference of convex (DC) programming}. We show the effectiveness of our approach through experiments on synthetic and benchmark datasets. Our approach performs better than the state of the art reject option classification approaches.

cs.LG↗

arXiv2014

An SVM Based Approach for Cardiac View Planning

We consider the problem of automatically prescribing oblique planes (short axis, 4 chamber and 2 chamber views) in Cardiac Magnetic Resonance Imaging (MRI). A concern with technologist-driven acquisitions of these planes is the quality and time taken for the total examination. We propose an automated solution incorporating anatomical features external to the cardiac region. The solution uses support vector machine regression models wherein complexity and feature selection are optimized using multi-objective genetic algorithms. Additionally, we examine the robustness of our approach by training our models on images with additive Rician-Gaussian mixtures at varying Signal to Noise (SNR) levels. Our approach has shown promising results, with an angular deviation of less than 15 degrees on 90% cases across oblique planes, measured in terms of average 6-fold cross validation performance -- this is generally within acceptable bounds of variation as specified by clinicians.

cs.LG↗

arXiv2014

A multi-instance learning algorithm based on a stacked ensemble of lazy learners

This document describes a novel learning algorithm that classifies "bags" of instances rather than individual instances. A bag is labeled positive if it contains at least one positive instance (which may or may not be specifically identified), and negative otherwise. This class of problems is known as multi-instance learning problems, and is useful in situations where the class label at an instance level may be unavailable or imprecise or difficult to obtain, or in situations where the problem is naturally posed as one of classifying instance groups. The algorithm described here is an ensemble-based method, wherein the members of the ensemble are lazy learning classifiers learnt using the Citation Nearest Neighbour method. Diversity among the ensemble members is achieved by optimizing their parameters using a multi-objective optimization method, with the objectives being to maximize Class 1 accuracy and minimize false positive rate. The method has been found to be effective on the Musk1 benchmark dataset.

cs.LG↗