Measuring Public Sentiment Towards Services Level in Online Forum using Naive Bayes Classifier and Word Cloud Andry Alamsyah, Faishal Nuruz Zuhri School of Economics and Business, Telkom University Bandung, Indonesia andrya@telkomuniversity.ac.id Abstract— In recent years, the rapid growth of the Internet has changed the way people interact globally. The internet usage is quite diverse, which one of them is a media to collect user generated content, including online review. Public sentiment is cumulative of people's arguments and opinions about the issues on public community. The sentiment could be expressed as positive or negative. One of online review type is the online forum, which contain specific topic that can affect people’s opinion in related discussions. This study aim to show how to measure polarity of public sentiment using machine learning principle. Even though this idea can be applied generally to any public domain, we choose business case study in telecommunication industry concerning their service level to the customer. We collect data from Kaskus, which is the most popular online forum in Indonesia. The interactions between users and their affection regarding the topic is measured. We use sentiment analysis based on naïve bayes classifier and word cloud approach in Bahasa to support the research objective. Keywords—sentiment analysis; topic detection; naive bayes classifier; word cloud. I. INTRODUCTION In recent years, the rapid growth of internet user has affected how people interact globally. The internet usage is quite diverse and multifunction, for example. in mass and personal communication, in business application, information sharing and contagion, in measuring public behavior, and many more. The internet has become the source of actual and universal information. Data shows that by March 2015, the number of active internet users in the world has reach 3.03 billion with the growth until 7.6 percent since 2014 [1]. There are many social media where people can exchange ideas and information. One of them is online discussion area, that have the advantage about topic separation comparing others type of social media. Online discussion forum raise up specific and thorough discussion, leads to comprehensive and representative public voice. One of the popular online forum in Indonesia is Kaskus. The number of its user has reached 8 million by April 2015. The interactions between users that create content are called as UGC (User Generated Content) [2]. UGC is so powerful for marketing, campaign, purchasing decision, and other application [3]. Public sentiment is an expression or the exchange of people's arguments and opinions about the issues on public community [4]. Sentiment analysis is involved in the study of opinion mining [5]. The simple forms to collect opinions are and retrieve data from social media. The opinions express public voice into two types; praises and complaints. When someone praise, it means positive polarity is formed. In other hand, when someone delivered a complaint, it means negative polarity is formed. There are wide variety of public sentiment classification other than positive or negative, such as; contextual, ambiguous, sarcasm, and the language used [6]. The product or service creation cannot be separated from consumer need. The typical buying process of consumer involves five stages, they are: problem recognition, information search, evaluation of alternatives, purchase decision, and post purchase behavior [7]. On buying process, business should concern to public sentiment. The whole process involves prospective buyers and consumer. The relations between prospective buyers and consumers can easily formed through social media. Mostly consumers post their experience after purchase the product, and then the prospective buyers searching information of the product that they are interested in. Sentiment analysis classification can be automated through machine learning using naïve bayes classifier and combined with topic detection using word cloud. Naïve Bayes Classifier is effective and accurate to classify public sentiments [8][9][10]. The topic detection is constructed using the word cloud to visualize, summarize and analyze the dominant word between positive and negative sentiment classification. The combination method between sentiment analysis and topic detection has been used in several research [11][12][13], but those research classify sentiment in English. Here we use Bahasa Indonesia which is different level of difficulties. By combining sentiment analysis and topic detection, a brand can learn market behavior. The information collected affect the organization in decision-making process, provide a view ahead, risk management, and decision evaluation. II. LITERATURE REVIEW A. Sentiment Analysis Sentiment analysis, or can be referred to the opinion mining, is the field of study of data mining which has the