Fast and Robust POS tagger for Arabic Tweets Using Agreement-based Bootstrapping

Fast and Robust POS tagger for Arabic Tweets Using Agreement-based Bootstrapping Fahad Albogamy, Allan Ramsay School of Computer Science The University of Manchester Manchester, M13 9PL, UK {fahad.albogamy, allan.ramsay}@manchester.ac.uk Abstract Part-of-Speech (POS) tagging is a key step in many NLP algorithms. However, tweets are difficult to POS tag because they are short, are not always written maintaining formal grammar and proper spelling, and abbreviations are often used to overcome their restricted lengths. Arabic tweets also show a further range of linguistic phenomena such as usage of different dialects, romanised Arabic and borrowing foreign words. In this paper, we present an evaluation and a detailed error analysis of state-of-the-art POS taggers for Arabic when applied to Arabic tweets. On the basis of this analysis, we combine normalisation and external knowledge to handle the domain noisiness and exploit bootstrapping to construct extra training data in order to improve POS tagging for Arabic tweets. Our results show significant improvements over the performance of a number of well-known taggers for Arabic. Keywords: Arabic tweets, POS tagging, Bootstrapping 1. Introduction The last few years have seen an enormous growth in the use of social networking platforms such as Twitter in the Arab World. A study prepared and published by Semiocast in 2012 has revealed that Arabic was the fastest growing language on Twitter in 2011. People post about their lives, share opinions on a variety of topics and discuss current is- sues. There are millions of tweets daily, yielding a corpus which is noisy and informal, but which is sometimes infor- mative. As a result, Twitter has become one of the most important social information mutual platforms. The nature of the text content of microblogs differs from traditional blogs. In Twitter, for example, a tweet is short and contains a maximum of 140 characters. Tweets also are not always written maintaining formal grammar and proper spelling. Slang and abbreviations are often used to overcome their restricted lengths (Java et al., 2007). POS tagging is an essential processing step in a wide range of high level text processing applications such as informa- tion extraction, machine translation and sentiment analysis (Barbosa and Feng, 2010). However, people working on Arabic tweets have tended to concentrate on low level lexi- cal relations which were used for shallow parsing and sen- timent analysis such as (Mourad and Darwish, 2013; El- Fishawy et al., 2014). They do not use the standard lin- guistic pipeline tools such as POS tagging which might en- able a richer linguistic analysis (Gimpel et al., 2011). The properties listed above of the microblogging domain make POS tagging on Twitter very different from its counterpart in more formal texts. It is an open question how well the features and techniques of NLP used on more well-formed data (e.g. in the newswire domain) will transfer to Twitter in order to understand and exploit tweets. Therefore, we experimentally evaluate the performance of state-of-the-art POS taggers for Modern Standard Arabic (MSA) on Arabic tweets. POS tagging accuracy drops from about …

Content truncated to 500 words