A general approach to natural language conversion

Natural language processing is a technique that includes both natural language understanding and natural language generation. Translating one natural language into another becomes complex due to structural difference, varieties of meanings, different forms of verbs etc. In this paper, a general algorithm has been developed which takes one natural language (i.e. English, German etc) as input and produces another natural language (i.e. Bengali, Japanese etc.) as output reserving the same expression. For understanding the input natural language, a smart-parse algorithm has been developed which unambiguously and efficiently generates a stack of the parse tree from the existing structure. After parsing, it incorporates itself with the knowledge base and dictionary and finally produces the corresponding targeted natural language. After fixing up the input and output natural language, it is possible to use this algorithm for translation after simple modification.