Report- ONe-class SVMs for document classification


1 Introduction

The article on study is titled "One-Class SVMs for Document Classification". It was written and published on 2001 in the journal of Machine Learning Research 2, a scientific journal that was founded on 2000, specialized in machine learning. The authors are Larry M.Manevitz and Malik Yousef. Manevitz is an  Associate Professor at the university of Haifa, Israel, specialized in Artificial Intelligence, Artificial Neural Network, Logic and Foundations of Mathematics, while Yousef is a PhD student in Machine Learning and Bioinformatics.
During this report, we will present the article from different views. First we will expose the general context and the existed solutions of the issue on study. After that, we will concentrate on the article's contributions. In order to prove their new methodology, the authors have done many experiments on existing famous data sets. We will be presenting the different results of these experiments later on the report. By the end, we will conclude by reminding the work done.

2 Context of the work
The context of the article is information retrieval. Manetvitz and Yousef propose a new method of document classification witch is based on SVM (Support Vector Machine) paradigm called Outlier methodology.
Classification, called also comparison module or similarity module, consists of comparing features issued from characterization  module. It assesses the degree of correspondence between two feature vectors, through a score.
SVM is a classifier derived from statistical learning theory. It is widely used in object detection and recognition, content-based image retrieval, text recognition, biometrics, speech recognition, etc. For SVM, training is based on both positive and negative examples.