Analysis of the Performance of the Damerau Levenshtein Distance and Peter Norvig Algorithms in Indonesian Spelling Correction

Abstract

Damerau Levenshtein Distance (DLD) and the Peter Norvig (PN) method are two types of algorithms used to perform spelling correction. The ease of application of these two algorithms has made many previous studies use them for word spelling correction. Therefore, this research aims to analyze and compare the performance of the two algorithms. This analysis only measures the accuracy and computation time required to perform spelling correction. From testing 500 wrong words and one edit, the DLD algorithm has an accuracy of 64.4% with a computation time of 67 seconds. Meanwhile, the PN method has an accuracy of 84% with a computation time of 0.21 seconds. For words with two edits, both have decreased. However, the PN method remains superior to DLD in both accuracy and computation time.