PERANCANGAN SISTEM KOREKSI OTOMATIS UJIAN ONLINE MULTIPLE CHOICE DAN ESSAY PADA KULIAH MIKROPROSESOR BERBASIS TF-IDF DAN VEKTOR SPACE MODEL DI STMIK ASIA MALANG

Tria Aprilianto, Jaenal Arifin

Sari


In this research will be developed an online-based system (website) that is able to conduct assessment or correction of the results of semester midterm or semester end automatically. About the exam given the system in the form of multiple choice and essay. Multiple choice selection process is done automatically by the system by automatically matching the results with the answer given by the lecturer, while the process of essay system essay giving a slightly different treatment that is by looking for similarity value / proximity to query or answer and alternative the answer given by the lecturer. The process of calculating similarity search between answers and queries is done by weighting TF-IDF (Term Frecuency - Invers Document Frecuency) and vector Space Modeling (Space Modeling Vector). TF-IDF is a term weighting method that is widely used as a benchmark against new weighting methods. In this method, the calculation of term t weight in a document is done by multiplying the value of Term Frequency with Inverse Document Frequency. The result of this Weighting is the similarity value of the document with the query. While the vector space modeling in this study is used to improve the value of similarity generated by weighting TF-IDF so that it is not possible to occur two the same highest value. The result of the method and modeling is then modified again so that it becomes a value that corresponds to the value (the desired answer rating weight). The process of preprocessing is done by the collection of documents obtained on the final exam of even semester microprocessor courses. From the results of the tests conducted then in this study it is concluded that the system built in www.ujianasia.net can work with the maximum and test algorithm done got 100% recall value, 85.4% precision and accuracy and similarity system against the answer given is 92.4%.

Kata Kunci


Website, Online Test, TF-IDF, Vector Space Model, Similarity, recall, precision

Teks Lengkap:

PDF

Referensi


Michael W. Berry and Malu Castellanos; Survey of Text Mining:Clustering, Classification, and Retrieval, Second Edition; 2007

RubĀ“ en Tous and Jaime Delgado; A Vector Space Model for semantic Similarity Calculation and OWL Ontology Alignment; 2009

Novi Safriadi dan Ari Wibowo; Uji Relevansi dan Performansi Sistem Temu Balik Informasi Pada Giggle Search Engine; 2011

Oka Karmayasa Dan Ida Bagus Mahendra; Implementasi Vector Space Model Dan Beberapa Notasi Metode Term Frequency Inverse Document Frequency (Tf-Idf) Pada Sistem Temu Kembali Informasi ;2012

Tristy Meinawati, Kodrat Iman Satoto, Oky Dwi Nurhayati; Perancangan Aplikasi Ujian Online Jurusan Sistem Komputer Universitas Diponegoro; 2013




DOI: https://doi.org/10.30736/jti.v2i2.70