Abstract-Case structure plays a vital role in grammatical structure of any language during language translation. This paper presents an in-depth analysis of Bangla locative case constructs based on UNL (Universal Networking Language) machine translation scheme. A set of analysis rules have been defined to convert various Bangla locative case sentences into UNL expressions that can later be converted to any native language using language independent deconversion rules. We have demonstrated five different analysis rules and illustrated how each of them can effectively convert Bangla sentences to UNL expressions Index Terms-Universal networking language (UNL), universal words (UWs), Bangla-UNL dictionary, morphological analysis, EnConverter (EnCo).
I. INTRODUCTIONUNL is a digital meta-language for describing, summarizing, refining and disseminating information in a machine independent and human languae neutral form, which represents information in the form of semantic networks with hypergraph. The hypergraph has formal English text realization as English is known to experts. It consists of Universal Words (UWs), UNL Relations and UNL attributes. An UW expresses the English equivalent meaning of the word along with some constraints lists and is to be used in creating UNL expression of output. UNL Relations are the building blocks of UNL expressions. The existence of UNL expressions relation is found in between two UWs of sentence. Relation between the words is drawn from a set of predefined relations. The UNL attributes are attached with UWs to provide additional information like tense, numbers etc. to complete the UNL expressions [1]. A set of analysis rules is to be used to generate UNL expressions from Bangla locative case sentences. The EnConverter [1], [2] is a language independent parse that provides synchronously a framework for morphological, syntactic and semantic analyses. EnConverter operates on the nodes of the Node-list through its windows. EnConverter analyses a sentence using the Word Dictionary, and enconversion Rules. It retrieves relevant dictionary entries from the word dictionary, operates on nodes in the Node-list by applying enconversion rules, and generates semantic networks of UNL by consulting the Knowledge Base. It generates UNL expressions from sentences of a native language using enconversion rules by finding the most Manuscript received August 8, 2013; revised December 12, 2013. Nawab Y. Ali and Ameer A. are with East West University, Dhaka, dmaa730@gmail.com In Fig. 1, aoj is the UNL relation which indicates "thing with attribute"; plc is another UNL relation which indicates "a place where an event occurs"; @entry and @ present are UNL attributes which indicate the main verb and tense information; and attribute @pl indicates number information.This organization of this paper is determined as following: The literature, which is related to the UNL structure and format of dictionary is illustrated in Section II. Analysis of different Bangla locative case sentences based on UNL structure an...