UTILIZATION OF SENTIMENT ANALYSIS USING THE DATA SCIENCE APPROACH TO IMPROVE CUSTOMER SATISFACTION
Ery Setiyawan Jullev Atmadji, Nanik Anita Mukhlisoh, Riski Ahmad Sulthoni

State Polytechnic Of Jember

Abstract

In 2018 the number of active Twitter social media users in Indonesia reached 22.8 million people and 4.1 billion tweets in 2016. Tweets or comments delivered can be used as a reference to determine and consider the business processes that will be carried out. However, the number of comments submitted is very large so that it has difficulty in analyzing the comment data. Besides the comment data on Twitter is an unstructured type of data so that if processing uses a relational database engine the results obtained are not optimal. To deal with these problems, a big data approach is needed in data extraction combined with the comment data processing model. This study uses a combination of big data in data processing and lexicon based to analyze customer comments. Data processing using big data especially with the NoSQL approach is very effective and efficient in conducting searches on unstructured data, this is because the search for big data is based on meta text rather than cardinality between data. While the lexicon based method is a method that depends on the completeness of the dictionary used. The purpose of this study is to analyze comments and share whether they have positive, negative, or neutral sentiments so that they can be used as parameters in decision making in an organization.

Keywords: Social Media, Twitter, Comment, Lexicon Based, Big Data, NoSQL

Topic: Others (Related to food and agriculture)

Link: https://ifory.id/abstract-plain/h6AxGbcgUyPM

Web Format | Corresponding Author (Ery Setiyawan Jullev Atmadji)