Combining Minimally-supervised Methods for Arabic Named Entity Recognition

Combining Minimally-supervised Methods for Arabic Named Entity Recognition Maha Althobaiti, Udo Kruschwitz, and Massimo Poesio School of Computer Science and Electronic Engineering University of Essex Colchester, UK {mjaltha, udo, poesio}@essex.ac.uk Abstract Supervised methods can achieve high perfor- mance on NLP tasks, such as Named En- tity Recognition (NER), but new annotations are required for every new domain and/or genre change. This has motivated research in minimally supervised methods such as semi- supervised learning and distant learning, but neither technique has yet achieved perfor- mance levels comparable to those of super- vised methods. Semi-supervised methods tend to have very high precision but compar- atively low recall, whereas distant learning tends to achieve higher recall but lower pre- cision. This complementarity suggests that better results may be obtained by combining the two types of minimally supervised meth- ods. In this paper we present a novel ap- proach to Arabic NER using a combination of semi-supervised and distant learning tech- niques. We trained a semi-supervised NER classifier and another one using distant learn- ing techniques, and then combined them using a variety of classifier combination schemes, including the Bayesian Classifier Combina- tion (BCC) procedure recently proposed for sentiment analysis. According to our results, the BCC model leads to an increase in per- formance of 8 percentage points over the best base classifiers. 1 Introduction Supervised learning techniques are very effective and widely used to solve many NLP problems, in- cluding NER (Sekine and others, 1998; Benajiba et al., 2007a; Darwish, 2013). The main disadvantage of supervised techniques, however, is the need for a large annotated corpus. Although a considerable amount of annotated data is available for many lan- guages, including Arabic (Zaghouani, 2014), chang- ing the domain or expanding the set of classes al- ways requires domain-specific experts and new an- notated data, both of which demand time and effort. Therefore, much of the current research on NER focuses on approaches that require minimal human intervention to export the named entity (NE) clas- sifiers to new domains and to expand NE classes (Nadeau, 2007; Nothman et al., 2013). Semi-supervised (Abney, 2010) and distant learn- ing approaches (Mintz et al., 2009; Nothman et al., 2013) are alternatives to supervised methods that do not require manually annotated data. These approaches have proved to be effective and easily adaptable to new NE types. However, the perfor- mance of such methods tends to be lower than that achieved with supervised methods (Althobaiti et al., 2013; Nadeau, 2007; Nothman et al., 2013). We propose combining these two minimally su- pervised methods in order to exploit their respective strengths and thereby obtain better results. Semi- supervised learning tends to be more precise than distant learning, which in turn leads to higher re- call than semi-supervised learning. In this work, we use various classifier combination schemes to combine the minimal supervision methods. Most previous studies have examined classifier combi- nation schemes to combine multiple supervised- learning systems (Florian et al., 2003; Saha and Ekbal, 2013), but this research is the …

Content truncated to 500 words