Efficient Error-Correcting Syntax Analysis for Recognition of Noisy Patterns,

Abstract : The use of a stochastic error-correcting parser for recognition of noisy and distorted patterns often results in a process of high accuracy, but with low efficiency. In this paper stochastic error-correcting parsers are formulated to include substitution, addition, and deletion errors. In order to make the syntax analysis practically feasible, a sequential classification algorithm with error-correcting capability is proposed. Considerable amount of parsing time can be saved by using the sequential algorithm. Computational results based on the classification experiments of noisy patterns for both non-sequential and sequential error-correcting parsers are presented.