Sentiment Analysis Online Tools: An Evaluation Study

Abstract

AbstractA sentiment analysis tool interprets text chats and assesses each opinion’s style, purpose, and feeling. The tool can better understand the context of users’ discussions, allowing the client service team to classify client feedback accurately. This is especially valuable for companies that actively address clients’ inquiries and complaints on social media, live chat, and email. Despite its vitality for business, there is still a challenge to decide the sentiment behind the content, especially for the Arabic language. Although most are not available for public usage, many sentiment analysis models and tools are developed in the literature. However, there is a lack of research identifying these tools’ practicality for the Arabic language. This paper investigates two pure online Arabic sentiment analysis tools by employing a sizeable Arabic dataset in the experiments. Prediction quality measurements were utilized to assess these tools. The yielded results recommended Sentest SA as a promised tool for detecting sentiment analysis polarity for the preprocessed Arabic social network contents.KeywordsSentiment analysisPolarityPrediction quality measurementsExperimental evaluation