<![CDATA[Customer complaints reveal product or service issues and can drive improvements to enhance satisfaction. Many companies use Twitter as a platform to interact with customers, making handling complaints through social media crucial for building a positive image and maintaining customer loyalty. IndiHome Regional 4 faces challenges in identifying main complaints due to a high volume of complaints on Twitter. Cluster analysis groups similar complaints, aiding the identification process. Text mining converts textual data into numerical format, streamlining complaint processing. Fuzzy Possibilistic C-Means Clustering, a fuzzy-based method, enables data membership across clusters with varying degrees of membership. By adopting relative (fuzzy) and absolute (possibilistic) membership, more accurate data placement is achieved. Data consists of IndiHome Regional 4 customer complaint tweets received via the Twitter channel "IndiHomeCare" from January to December 2022. The clustering process formed 4 clusters based on the smallest Extended Xie-Beni Index value, tested with different cluster numbers (3-7). Witel Yogyakarta had the highest members and complaints in each cluster, while Witel Kudus had the lowest. Word Cloud analysis revealed main complaints in each cluster, including WiFi-related subscription costs, internet disruptions, customer service issues, and slow connections.]]>
Copyrights © 2025
