AMBRA: A Ranking Approach to Temporal Text Classification

Proceedings of the 9th International Workshop on Semantic Evaluation (SemEval 2015), pages 851–855, Denver, Colorado, June 4-5, 2015. c⃝2015 Association for Computational Linguistics AMBRA: A Ranking Approach to Temporal Text Classification Marcos Zampieri1,2, Alina Maria Ciobanu3, Vlad Niculae4, Liviu P. Dinu3 Saarland University, Germany1 German Research Center for Artificial Intelligence (DFKI), Germany2 Center for Computational Linguistics, University of Bucharest, Romania3 Deptartment of Computer Science, Cornell University, USA4 marcos.zampieri@uni-saarland.de; alina.ciobanu@my.fmi.unibuc.ro; vn66@cornell.edu; ldinu@fmi.unibuc.ro; Abstract This paper describes the AMBRA system, entered in the SemEval-2015 Task 7: ‘Di- achronic Text Evaluation’ subtasks one and two, which consist of predicting the date when a text was originally written. The task is valuable for applications in digital humani- ties, information systems, and historical lin- guistics. The novelty of this shared task con- sists of incorporating label uncertainty by as- signing an interval within which the document was written, rather than assigning a clear time marker to each training document. To deal with non-linear effects and variable degrees of uncertainty, we reduce the problem to pair- wise comparisons of the form is Document A older than Document B?, and propose a non- parametric way to transform the ordinal output into time intervals. 1 Introduction Temporal text classification consists of learning to automatically predict the publication date of docu- ments, by using the information contained in their textual content. The task finds uses in fields as var- ied as digital humanities, where many texts have are unidentified or controversial publication dates, in- formation retrieval (Dakka et al., 2012), where tem- poral constraints can improve relevance, and his- torical linguistics, where the interpretation of the learned models can confirm and reveal insights. From a technical point of view, the task is usu- ally tackled either as regression or, more commonly, as a single-label multi-class problem, with classes defined as time intervals such as months, years, decades or centuries. The regression approach as- sumes that precise timestamps are uniformly avail- able for each document, which is suitable for cases of social media documents (Preotiuc-Pietro, 2014), but less suitable for documents surrounded by more uncertainty. Multi-class classification, on the other hand, suffers from a coarseness tradeoff: using coarser classes is less informative, and using finer classes reduces the number of training instances in each class, making the problem more difficult. Fur- thermore, with a multi-class formulation, the tempo- ral relationship between classes is lost. The ‘Diachronic Text Evaluation’ subtasks one and two from SemEval-2015 are formulated simi- larly to a multi-class problem, where each document is assigned to an interval such as 1976-1982. To accommodate such labels, we propose an approach based on pairwise comparisons. We train a classi- fier to learn which document out of a pair is older and which is newer. If two documents come from overlapping intervals, then their order cannot be de- termined with certainty, so the pair is not used in training. We use the property of linear models to ex- tend a set of pairwise decisions into a ranking of test documents (Joachims, …

Content truncated to 500 words