Indonesian Journal of Electrical Engineering and Computer Science Vol. 28, No. 2, November 2022, pp. 890~897 ISSN: 2502-4752, DOI: 10.11591/ijeecs.v28.i2.pp890-897  890 Journal homepage: http://ijeecs.iaescore.com Healthcare assessment for beauty centers using hybrid sentiment analysis Abeer Khalid Al-Mashhadany 1 , Ahmed T. Sadiq 2 , Sura Mazin Ali 3 , Amjed Abbas Ahmed 4 1 Department of Computer Science, Al-Nahrain University, Baghdad, Iraq 2 Department of Computer Science, University of Technology, Baghdad, Iraq 3 Politiical Sciences College, AlMustansiriyah University, Baghdad, Iraq 4 Imam AlKadhum College, Baghdad, Iraq Article Info ABSTRACT Article history: Received Mar 23, 2022 Revised Jul 12, 2022 Accepted Aug 11, 2022 Because of COVID-19, healthcare became the first interesting domain at the world. Here, comes the role of researchers to do what they can to guide people. Nowadays, the most wanted field is beauty industry. It achieved large market. And the estimation is toward the growing. Researchers can give advice to prevent unhealthy causes in this field. They can apply sentiment analysis methods to make decision whether a Beauty center is healthy or unhealthy. This work develops an improved method of sentiment analysis to classify the beauty centers in Iraq into healthy and unhealthy classes. Researchers used comments of beauty centers’ Facebooks to perform the assessment. The methodologies encompass the two approaches lexicon-based and machine-learning-based. Three machine learning mechanisms had been applied; rough set theory, naïve bayes, and k-nearest neighbors. It will be shown that rough set theory is the best compared with the others two. Rough set theory achieved 95.2%, while Naïve Bayes achieved 87.5% and k-nearest neighbors achieved 78%. Keywords: Assessment K-nearest neighbors Naïve Bayes Rough set theory Sentiment analysis Social media Text mining This is an open access article under the CC BY-SA license. Corresponding Author: Abeer Khalid Al-Mashhadany Department of Computer Science, Al-Nahrain University Baghdad, Iraq Email: aabeeeeraa@yahoo.com 1. INTRODUCTION The world after 2019 differs from its before. Health care became the first interesting domain at all countries, and for everybody. Many and many topics missing health care researches. The modern era directs people to be consumers of what is available to them. The motive is to achieve business deals for the benefit of kapitals in the world. People are drifting strongly toward the current, because of their desire to imitate, or in other words walk with the herd. Here comes the role of researchers to do what they can to guide people. The most wanted field is beauty industry. Its market size in 2020 was $483 billion. It is estimated to achieve $716 billion end 2025 [1], [2]. Researchers can give advice to prevent unhealthy causes in this field. They are applying text analysis approaches which are useful to make decision; healthy, or unhealthy. Text analysis provides approaches to describe and interpret text’s characteristics. As one of the common text analysis applications is document classification. Text may be structured-text; found in an organized form. Really, most examples of texts are unstructured-text such as chat-rooms. Analyzing structured-text is easier than analyzing unstructured-text. Today, sentiment mining from unstructured-text is the commonly used approach to classify text into negative or positive. Negative class gives advice that something is bad, so the decision will be leave it. While positive class gives advice that something is good, so the decision will be use it [3]-[5].