Performance Estimation of Stochastic Gradient Descent Classifier Using Confusion Matrix and ROC Curve
Published: 2019
Author(s) Name: Puneet Kumar and Somil Jain |
Author(s) Affiliation: Assist. Prof., Dept. of Comp. Scie. & Engg., SET, Mody Univ. of Science and Tech., Rajasthan, India.
Locked
Subscribed
Available for All
Abstract
Whether an email is a spam or not? Whether the image containing the image of cat or not? Such type of questions in machine learning can be answered only by using process of classification. The uses of classification under the umbrella of supervised learning are inevitably important. So in the voyage of classification, a classifier need to be used which is responsible for apt classification or categorization of a data into its valid category. Further, the performance of the classifier chosen for the course of action will be analysed in order to get insights about its accuracy. Hence, in this attempt, Stochastic Gradient Descent binary classifier will be chosen for classification based prediction of an image carrying handwritten digits. Afterwards, the performance will be analysed by using cross validation, confusion matrix and ROC curve analysis.
Keywords: Cross validation, Machine learning, Precision, Recall, Sensitivity.
View PDF