Fulfillment and Responsiveness on Online Travel Agencies Using Multiclass Classification
Nadya Nurul Sepani Putri, Andry Alamsyah, Sri Widiyanesti · 2020
In recent years, Online Travel Agencies (OTA) is widely used by people due to its simplicity and efficiency. Tight competition between industries makes companies must pay attention to the quality of their services since it is capable to enhance customer satisfaction. To evaluate its service, the company needs to comprehend their position as OTA providers. Users' opinion in social media is essential for recognizing the company performances. In this case, sentiment analysis and multiclass classification methods help the company to understand their service quality specifically. As a case study, we use the most popular Online Travel Agencies (OTA) providers in Indonesia: Traveloka, Tiket.com, and Pegipegi. Based on our criterion, we examine the fulfillment and responsiveness dimensions of these three OTA providers. We apply Naïve Bayes Classifier (NB) model to classify users' opinions. This model has accuracy around 75-85% for the three OTA providers. The result reveals that Pegipegi obtains better service quality with 57% positive and 43% negative sentiment than Traveloka and Tiket.com with 56% positive and 44% negative. The overall result shows general topics of fulfillment and responsiveness dimensions are related to the ticket availability and customer service performances.