Prototype of Online Examination on MoLearn Applications Using Text Similarity to Detect Plagiarism

The existence of mobile learning is a logical consequence of information technology advancement in education. In the previous research, the produced mobile learning application called MoLearn can only handle automatic assessments of online exam with multiple choice questions. Currently, MoLearn has not been able to fully meet needs of school on the learning evaluation of essay questions. Automatic assessment of essay answers is more difficult to match with key answers, but the reality is still a lot of exams use the type of essay questions. Therefore, the solution offered is the creation of online exam prototype that can perform essay assessment using text similarity method. The text similarity methods that can be used to assess the correctness level of answer and plagiarism level of answers among examinees are cosine similarity, Smith-Waterman, and Latent Semantic Analysis (LSA) algorithms. The results show that this prototype can give MoLearn application development proposal to handle automatic assessment of essay questions. In addition, this prototype can also display the level of plagiarism answers among students as an effort to promote the value of honesty.