Error Detection of Personalized English Isolated-Word Using Support Vector Machine
Tiong Sieh Kiong, Abu Bakar Hasa, Johnny Koh Siaw P, David Yap Fook W · Trends in Applied Sciences Research · 2012
A better understanding on word classification could lead to a better detection and correction technique. In this study, a new features representation technique is used to represent the machine-printed English word. Subsequently, a well-known classification type of artificial intelligent algorithm namely Support Vector Machine (SVM) is used to evaluate those features under two class types of words with proper segregation of correct and erroneous words in two data sets. Our proposed model shows good performance in error detection and is superior when compared with neural networks, Hamming distance or minimum edit distance technique; with further improvement in sight.