With the recent development of utf-8 encoding format, lot of focus has been shifted toward development of artificially intelligent machines that can handle Hindi language text. Lexical, Syntactic and Semantic analyses are the essential steps in processing applications involving any natural language. These act as major pre-processing tasks in applications like information extraction, text summarization, machine translation etc. Moreover these are also helpful to obtain better results as efficiency of such machines depends mostly upon percentage of ambiguities that had been resolved in Lexical, Syntax and Semantic processing phases of natural language itself. In this paper we present a method to perform syntax analysis of Hindi sentences via probabilistic parsing technique using CYK algorithm to build parse table and a method to detect semantical correctness/incorrectness of Hindi sentence by using morphological information of words provided by morphological analyzer tool of IIIT Hyderabad, India. Our work achieved overall accuracy of approximately 80% in resolving ambiguities from Hindi sentences at syntax processing stage and accuracy of approximately 89% in detecting semantical errors in semantic analysis phase (provided that sentence was syntactically correct).
