This paper presents a formal analysis of popular text classification methods, focusing on their loss functions whose minimization is essential to the optimization of those methods, and whose decomposition into the trainingset loss and the model complexity enables cross-method comparisons on a common basis from an optimization point of view. Those methods include Support Vector Machines, Linear ...

Exploring digital collections to find information relevant to a user’s interests is a challenging task. Algorithms designed to solve this relevant information problem base their relevance computations on user profiles in which representations of the users’ interests are maintained. This paper presents a new method, based on the classical Rocchio algorithm for text categorization, able to discov...

Exploring digital collections to find information relevant to a user’s interests is a challenging task. Algorithms designed to solve this relevant information problem base their relevance computations on user profiles in which representations of the users’ interests are maintained. This article presents a new method, based on the classic Rocchio algorithm for text categorization, able to discov...

Most of the approaches in multi-view categorization use early fusion, late fusion or co-training strategies. We propose here a novel classification method that is able to efficiently capture the interactions across the different modes. This method is a multi-modal extension of the Rocchio classification algorithm – very popular in the Information Retrieval community. The extension consists of s...

This paper explores the use of a statistical technique known as density estimation to potentially improve the results of text categorization systems which label documents by computing similarities between documents and categories. In addition to potentially improving a system's overall accuracy, density estimation converts similarity scores to probabilities. These probabilities provide con denc...

We describe a new family of topic-ranking algorithms for multi-labeled documents. The motivation for the algorithms stem from recent advances in online learning algorithms. The algorithms are simple to implement and are also time and memory efficient. We provide a unified analysis of the family of algorithms in the mistake bound model. We then discuss experiments with the proposed family of top...

We built a filtering system YFILTER this year, which we used for experiments on profile updating and thresholds setting. Our focus is using incremental Rocchio for introducing new query terms and term weighting. Although 1, 0.5, 0.25 is a widely used Rocchio ratio for query expansion based on relevance feedback, we found that the optimal setting for information filtering is corpus and profile d...