Improvement of Thai error correction system by memetic algorithm

Abstract

This paper presents an efficient technique for improving the efficiency of Thai error correction system by
using memetic algorithm. In this paper, the token passing algorithm is used for constructing word graph and
the language model is used checking the correct sentence. The correction process starts with word graph
construction from token passing algorithm, then the correct sentence are searched by memetic algorithm
with the best fitness function from language model. For a long sentence from the token passing algorithm, a
search space is very huge which can be resolved by using memetic algorithm. The memetic algorithm is
used for searching the correct sentence in order to reduce the analysis time. The performance of the
proposed method are evaluated and compared to the full search and genetic algorithm. From the experimental results show that the proposed method performs very well and yields better performance more
than the compared method. The proposed method can search the best sentence accurately and quickly.