Comparison of cosine similarity and k-NN for automated essays scoring

Ewees, Ahmed A.; Eisa, Mohamed; Refaat, Mark

doi:10.17148/ijarcce.2014.31202

Cited by 12 publications

(9 citation statements)

References 16 publications

(15 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Several methods and techniques for measuring the lexical similarity in the retrieval of Arab information (IR) were studied, concluded the cosine similarity as better measurement compared to other techniques and methods [36]. In another study [37], the Cosine similarity was used to study the classification of Arabic texts, whereby VSM , k-Nearest Neighbors, Naïve Bayes, classification tree, and Neural Network are the methods employed in this study [38]. The cosine similarity was also compared with the algoritm called K-NN from the LSA method and it runs automatic to record Arabic articles [38].…”

Section: B) Statement-based Similaritymentioning

confidence: 99%

“…In another study [37], the Cosine similarity was used to study the classification of Arabic texts, whereby VSM , k-Nearest Neighbors, Naïve Bayes, classification tree, and Neural Network are the methods employed in this study [38]. The cosine similarity was also compared with the algoritm called K-NN from the LSA method and it runs automatic to record Arabic articles [38].…”

Section: B) Statement-based Similaritymentioning

confidence: 99%

See 1 more Smart Citation

Arabic Language Documents’ Similarity and its Challenges (A Review)

Saleh¹,

Fadzli²,

El-Ebiary³

2020

IJETT

View full text Add to dashboard Cite

Measuring the similarity of documents is of great importance and has influences on many areas and subjects such as Information Retrieval (IR), redrafting discovered, classified documents, and conversational agent programs. Recently, several scientific studies have been interested in implementing the similarity of documents in Arabic language applications. In this study, we review most of the techniques used in measuring the similarity between documents in Arabic; these techniques are classified into three types, mainly lexical, semantic, and finally hybrid. Hence, most of the techniques and methods adopted to measure the similarity between documents were reviewed to determine the most appropriate ones to measure the similarity in Arabic. However, the measurement of the percentage of similarity in documents in the Arabic language represents a major challenge faced by most of the researchers due to the nature of the Arabic language, such as complex morphological processes, ambiguity, and a general lack of resources.

show abstract

Section: B) Statement-based Similaritymentioning

confidence: 99%

Section: B) Statement-based Similaritymentioning

confidence: 99%

Arabic Language Documents’ Similarity and its Challenges (A Review)

Saleh¹,

Fadzli²,

El-Ebiary³

2020

IJETT

View full text Add to dashboard Cite

show abstract

“…Selain metode yang disebutkan diatas terdapat metode lain yang dapat digunakan untuk melakukan penilaian jawaban esai otomatis yaitu dengan metode searching text similarity [8]. Metode digunakan untuk menghitung kesamaan atau kemiripan dari dua buah dokumen.…”

Section: Pendahuluanunclassified

E-learning Automated Essay Scoring System Menggunakan Metode Searching Text Similarity Matching Text

Ahmad

2019

jpe

View full text Add to dashboard Cite

E-learning is an information technology that has been widely applied to the field of education in the form of websites because it can be accessed indefinitely and place. The development of E-learrning is logical because it follows the age of putting forward information and communication technology. Conventional learning is especially an evaluation of essay form learning, where teachers check students' essay answers that require more time and effort. Automated Essay Scoring (AES) is one of the features of Elearning that can facilitate the teacher in matching the answer of student essay with answer key. Text similarity is a method of matching text that is recall word based on specified condition (term). The language used in AES is Indonesian which has a different word morphology with other languages. The results obtained in this study is the Error rate using the term frequency (tf) is lower with high accuracy.

show abstract

“…Sentence Feature Calculation: Each sentence is given a score, which serves as a good measure of the sentence by using a set of specific features. Each preset feature score takes a value ranging from (1,0), the following set of features will be used:  Frequency Feature: frequency of word play a crucial role, to decide the importance of any word or sentence in a given document [12,17,18]. In our method, the weight of the sentence is calculated on the basis of the frequency of the term or synonyms and frequency of relations by calculating the average value of the frequency of the term in each sentence as well as the synonyms.…”

Section: The Processing Stagementioning

confidence: 99%

A Proposed Method for Summarizing Arabic Single Document

Bialy¹,

Ewees²,

Elgamal³

2018

IJCA

View full text Add to dashboard Cite

This paper proposes an automatic text summarization method, which is considered as a selective process for the most important information in the original text. It could be divided into two types extractive and abstractive. In this study, a system for single documents text summarization is introduced to be used for Arabic text that rely on extractive method. According to this, we will go three stages, which are preprocessing phase, scoring of sentence, and summery generation. The pre-processing phase starts by removing punctuation marks, stop words, unifies synonyms as well as stemming words to obtain root form. Then it measures every sentence according to a collection of features in order to get the sentences with a higher score to be included in the final summary. The system has been evaluated by comparing between manual and automatic summarizations and some measurements are used especially Rouge measure. Manual summarize is done by two human experts to check the summaries' quality in terms of the general form, content, coherence of the phrases, lack of elaboration, repetition, and completeness of the meaning. The final results proved that the proposed method achieved the higher performance than other systems.

show abstract

Comparison of cosine similarity and k-NN for automated essays scoring

Cited by 12 publications

References 16 publications

Arabic Language Documents’ Similarity and its Challenges (A Review)

Arabic Language Documents’ Similarity and its Challenges (A Review)

E-learning Automated Essay Scoring System Menggunakan Metode Searching Text Similarity Matching Text

A Proposed Method for Summarizing Arabic Single Document

Contact Info

Product

Resources

About