TY  - JOUR
T1  - A Comparative Analysis of Euclidean Distance and Cosine Similarity
Measure for Automated Essay-Type Grading
AU - E. Oduntan, Odunayo. AU - Obe, Olumide O. AU - S. Falohun, Adeleye AU - A. Adeyanju, Ibrahim 
JO  - Journal of Engineering and Applied Sciences
VL  - 13
IS  - 11
SP  - 4198
EP  - 4204
PY  - 2018
DA  - 2001/08/19
SN  - 1816-949x
DO  - jeasci.2018.4198.4204
UR  - https://makhillpublications.co/view-article.php?doi=jeasci.2018.4198.4204
KW  - reduced vector
KW  -modified principal component algorithm
KW  -automated essay-type grading system
KW  -Euclidean distance measure
KW  -cosine similarity measure
KW  -Evaluation
AB  - Evaluation of student’s performance is inevitable in any educational setting, allocating scores to
student’s response is a function of how close the answer supplied to the question is to expected answer. This
study delves into analyzing the effectiveness of cosine similarity measure and Euclidean distance which are
both used in similarity measures for Automated Essay Type Grading System (AETGS). AETGS involves
transcription of the contents of the marking schemes into electronic form to derive a txt file extension using text
editor while student’s answers assumed txt format. The inherent stopwords and stemming in the txt document
were pre-processed to address morphological variations using standard stopwords list and porters stemmer
algorithm, respectively. N-gram terms were derived for each student’s response and the Marking Schemes (MS)
using the vector space model. A Document Term Matrix (DTM) was generated with N-gram terms of MS and
students response representing columns and rows, respectively. Modified principal component analysis
algorithm was used to reduce the sparseness of the DTM to obtain a vector representation of the student’s
answers and the marking scheme. The reduced vector representation of the student’s answers was graded
according to the mark assigned to each question in the marking scheme using cosine similarity measure and
the Euclidean distance measure. The developed Automated Essay-Type Grading System (AETGS) was
implemented in Matrix Laboratory 8.1 (R2013a). The effect of the similarity measures on the developed system
was performed using Pearson Correlation coefficient of two courses: CMP401-Organization of Programming
Languages and CMP205-Operating System I. The result showed that cosine similarity measure has a high
positive correlation than the Euclidean distance.
ER  -