A Naïve Bayes Approach to Classifying Topics in Suicide Notes

Este paper bastante interessante sobre Text Mining (Mineração sobre bases textuais) trata de uma análise sobre cartas de suicídio e foi apresentado na I2B2 Challenge on Sentiment Classification.

O abstract traz informações relevantes sobre o método de trabalho e o resultado, porém; por mais doentio que possa parecer em um primeiro momento devido a morbidade do título; a iniciativa é amplamente válida para estudos relacionados a classificação e identificação de padrões de características que podem ajudar estudos psiquátricos, médicos, e até famacológicos na busca de atenuação desse tipo de comportamento humano.

A Naïve Bayes Approach to Classifying Topics in Suicide Notes

A Naïve Bayes Approach to Classifying Topics in Suicide Notes

Authors: Irena Spasic, Pete Burnap, Mark Greenwood and Michael Arribas-Ayllon Publication Date: 30 Jan 2012 Journal: Biomedical Informatics Insights Citation: Biomedical Informatics Insights 2012:5 (Suppl. 1) 87-97

Abstract The authors present a system developed for the 2011 i2b2 Challenge on Sentiment Classification, whose aim was to automatically classify sentences in suicide notes using a scheme of 15 topics, mostly emotions. The system combines machine learning with a rule-based methodology. The features used to represent a problem were based on lexico–semantic properties of individual words in addition to regular expressions used to represent patterns of word usage across different topics. A naïve Bayes classifier was trained using the features extracted from the training data consisting of 600 manually annotated suicide notes. Classification was then performed using the naïve Bayes classifier as well as a set of pattern–matching rules. The classification performance was evaluated against a manually prepared gold standard consisting of 300 suicide notes, in which 1,091 out of a total of 2,037 sentences were associated with a total of 1,272 annotations. The competing systems were ranked using the micro-averaged F-measure as the primary evaluation metric. Our system achieved the F-measure of 53% (with 55% precision and 52% recall), which was significantly better than the average performance of 48.75% achieved by the 26 participating teams.