LGJun 28, 2014
Comparison of SVM Optimization Techniques in the Primal
arXiv:1406.7429v1
Originality Synthesis-oriented
AI Analysis
This work is incremental, addressing efficiency improvements for SVM practitioners in text classification tasks.
This paper compared three optimization techniques for primal SVM formulations, finding that one technique achieved a 2.5% higher accuracy on the Sentiment Analysis on Movie Reviews dataset.
This paper examines the efficacy of different optimization techniques in a primal formulation of a support vector machine (SVM). Three main techniques are compared. The dataset used to compare all three techniques was the Sentiment Analysis on Movie Reviews dataset, from kaggle.com.