<![CDATA[The development of information technology has driven a transformation in the learning evaluation system, particularly through the implementation of web-based online exams. However, the assessment of essay answers remains a challenge because it is generally done manually, which is time-consuming and potentially subjectivist. This study aims to implement the Winnowing Algorithm in a web-based automated essay answer assessment system at Alifa University, Padang. The Winnowing Algorithm is used to measure the level of text similarity between student answers and answer keys using document fingerprinting techniques. The research method used is an experiment with preprocessing stages, n-gram formation, hash calculation using rolling hash, window formation, and fingerprint selection. The data used are student essay answers and answer keys determined by the lecturer. The results show that the system is able to calculate the level of similarity automatically with similarity values ranging from 50% to 90%. Examples of manual calculations produce hash values of 3192 and 3262 for the two sets of texts tested. Thus, the Winnowing Algorithm can be an effective solution in the development of a web-based automated essay answer assessment system.]]>
Copyrights © 2026
