Porter Stemmer and Cosine Similarity for Automated Essay Assessment

Nunik Destria Arianti, Mohamad Irfan, Undang Syaripudin, Dina Mariana, Neny Rosmawarni, Dian Sa’adillah Maylawati

2019 · 12 citations · 25 references

Concepts

Abstract

Essay assignments are one of the effective ways to find out the understanding of students about the material that has been studied in both high school and university. Of course, the essay is different from multiple choice questions which are easily automated to assess, while the essay has a text narrative that needs to be known and maintained in its meaning. This article aims to make an automatic assessment of essay assignments by using text mining techniques to maintain the meaning in the text. One process in text mining is a stemming process that returns words to their basic word form. The stemming process is important because stemming words are verbs that store the meaning of the text. This article discusses the Porter and Cosine Similarity Algorithms in the stemming process. Based on the experimental results of 35 essay assignments, the automatic assessment of the essay has an accuracy rate of around 97% compared to the results of the expert assessment, with the difference in value between automatic assessment and expert judgment 2.62 points.

References

25