Computer Based Test Using the Fisher-Yates Shuffle and Smith Waterman Algorithm

Abstract

Tests are used to determine a person’s level of understanding of a subject. The inhibiting factors in tests are less varied questions, questions with insufficient difficulty, subjective assessments, and the length of time in their correction. This research aimed to develop a Computer Based Test (CBT) application. The type of questions in this CBT are multiple choice and essays. This CBT employs categorization of questions, randomization of the questions, and automatic assessment. Questions were categorized manually based on Bloom’s Taxonomy of a lecture. Then the randomization process was carried out using the Fisher-Yates Shuffle algorithm for each question category. The Smith Waterman algorithm was used to automatically assess the essay-type questions. The steps of the Smith Waterman algorithm were preprocessing, data comparison using Smith Waterman, and percentage similarities conversion to test scores. The results of the study showed that the CBT application was able to randomize questions using the Fisher-Yates Shuffle algorithm and automatically assess answers using the Smith Waterman algorithm. RMSE was used to measure of the accuracy of the Smith Waterman algorithm: a value of 1.86 was obtained. Keywords: Computer based test, assessment, Fisher-Yates Shuffle, Smith Waterman