Article Preview
Top1. Introduction
Internet has evolved from monolingual into a multilingual World Wide Web for more than 1000 languages shown in (Ali et al., 2012). The Internet usage by the vast number of users now varies from primarily academic purposes to widespread commercial, leisure, education, entertainment etc. purposes. In Internet, vast knowledge and information resources in different languages are scattered all over the world and remain mostly inaccessible due to non-machine representation and language barrier in information distribution as shown in (Ali et al., 2008). It has been showed that there is a great need to translate web pages and electronic mail messages into the native language for overcoming the language barrier (Kumar et al., 2011). Considering this issue, Universal Networking Language (UNL) has been developed by the United Nations University/Institute of Advanced Studies (UNU/IAS) to convert any language to other language presented in (Ali et al., 2013). The UNL system is introduced most of the languages like Arabic, Spanish, Japanese, Chinese, Russian and many more as shown in (Choudhury et al., 2005). A UNL based transfer system needs two components: one for conversion from source language to UNL expression and another for UNL expression to target language. The UNL system has 46 semantic relations and 87 attributes to express the semantic meaning of a sentence of a source language demonstrated by (Uchida et al., 1993). With this increasing pressure of providing information access without language and cultural barriers, Internet today has to deal with immense complexity of multilinguality.
Bangla is the 4th widely spoken language in the world with more than 250 million speakers, most of whom live in Bangladesh and the Indian state of West Bengal which is evidenced (Ali et al. 2008). Bangla text is converted to UNL expressions by the EnConverter of Bangla language. This UNL expression is converted to any other language that has its own DeConverter. This will indeed help to develop a multilingual machine translation system for Bangla language.
A statistical Bangla to English translation engine using only simple Bangla sentences that contain a subject, an object and a verb is developed by (Uddin et al., 2004). A low-cost English to Bangla (E2B)-ANUBAD translating, English text into Bangla text using both rule-based and transformation-based MT schemes along with three-level of parsing is shown in (Saha, 2005). MT Bangla dictionaries that address the organization, contents and details of the information have been developed by (Ali et al., 2002). The research has been done for morphological analysis of Bangla words for UNL by (Choudhury et al., 2005), parsing methodology for Bangla sentences by (Asaduzzaman et al., 2003), and dictionary development of Bangla words by (Ali et al., 2002; Islam, 2009). The suffix, prefix and inflexions of Bangla language are detailed in (Asaduzzaman et al., 2008; Khairunnahar, 2008).