A Phrase-Based Machine Translation from English to Bangla Using Rule-Based Approach

In this paper, a model of transfer architecture has been proposed which represents a Rule-Based Approach. This approach relies on the fuzzy rules. It is a tense and phrase based English to Bangla transfer system. This article represents a knowledge-based technique with a set of data. A rough set technique is used in knowledge representation system for language translation. This technique is used to categorize each English sentence to a particular group using attributes and organized in a pattern. The pattern arranges according to the rules then the system will produce the target sentence Bangla. The whole procedure completes with 6 steps: 1) Collect data 2) Tokenized by word 3) Arrange according to rules 4) Morphological Analyze 5) Reconstruct Bangla sentence using appropriate rule 6) Target sentence. Comparing the experimental result with Google translator, it has been found that the model translation system provides higher accuracy then comparing translator.