Luis Alfonso Ureña López

Adaptive Selection of Base Classifiers in One-Against-All Learning for Large Multi-labeled Collections

In this paper we present the problem found when studying an automated text categorization system for a collection of High Energy Physics (HEP) papers, which shows a very large number of possible classes (over 1,000) with highly imbalanced distribution. The collection is introduced to the scientific community and its imbalance is studied applying a new indicator: the inner imbalance degree. The one-against-all approach is used to perform multi-label assignment using Support Vector Machines. Over-weighting of positive samples and S-Cut thresholding is compared to an approach to automatically select a classifier for each class from a set of candidates. We also found that it is possible to reduce computational cost of the classification task by discarding classes for which classifiers cannot be trained successfully.

Domain Adaptation of Polarity Lexicon combining Term Frequency and Bootstrapping

In this paper we study several approaches to adapting a polarity lexicon to a specific domain. On the one hand, the domain adaptation using Term Frequency (TF) and on the other hand, the domain adaptation using pattern matching with a BootStrapping algorithm (BS). Both methods are corpus based and start with the same polarity lexicon, but the first one requires an annotated collection of documents while the second one only needs a corpus where it looks for linguistic patterns. The performance of both methods overcomes the baseline system using the general polarity lexicon iSOL. However, although the TF approach achieves very promising results, the BS strategy does not give as much improvement as we expected. For this reason, we have combined both methods in order to take advantage of the positive aspects of each one. With this new approach the results obtained are even better that those with the systems applied individually. Actually, we have achieved a significant improvement of 11.50% (in terms of accuracy) in the polarity classification of the movie reviews with respect to the results achieved with the general purpose lexicon iSOL.

Buscador Semántico Biomédico

Este trabajo esta parcialmente subvencionado por el proyecto REDES (TIN2015-65136-C2-1-R) del MICINN del Gobierno de Espana.