Similarity-based Learning Approaches to Arabic Sentiment Analysis
Journal of Environmental Science Computer Science and Engineering & Technology · 2019
This paper investigates opinion detection and sentiment analysis (SA) to automatic classifying Arabic comments and placing text documents into proper category positive or negative based on their contents.In this paper K-Nearest Neighbor (K-NN) classifier is used to determine the polarity of the document and is studied the performance of K-NN classification when applying the different distance measures.The K-NN algorithm is based on learning by comparing a given test documents with training documents.Measuring the similarity between documents is presented also in the paper; the three distance metrics (Euclidean distance, Cosine similarity and jaccard distance) are presented.The effect of these three measures is performed on three available resources (dataset) composed of Arabic language.The Evaluation parameters such as accuracy, recall and precision are measured the experimental results give that the performance of the jaccard measure with k =15 is better than other measures for Dataset.