Putri, Cahya Buana (2019) Text Analytics pada Maskapai Penerbangan Indonesia di Twitter Menggunakan Metode Support Vector Machine. Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
06211540000112-Undergraduate_Theses.pdf Download (3MB) | Preview |
Abstract
Seiring dengan perkembangan teknologi dan media online yang cepat, komentar dan komplain pelanggan dapat dilihat dari berbagai media sosial, termasuk Twitter yang sering digunakan sebagai media berkomunikasi dengan pelanggan oleh pelaku bisnis termasuk di bidang transportasi udara. Maskapai penerbangan Indonesia yang digunakan dalam penelitian ini adalah AirAsia dan LionAir. Penelitian ini menerapkan pendekatan text mining, dimana setiap tweet bersentimen akan dilakukan klasifikasi menggunakan Support Vector Machine dengan Synthetic Minority Over-sampling Technique (SMOTE) untuk menangani kasus imbalanced data dan n-gram sebagai tahap tokenizing pada pre processing. Hasil klasifikasi dievaluasi menggunakan nilai AUC (Area Under Curve). Model klasifikasi terpilih adalah hasil klasifikasi yang memberikan nilai AUC paling tinggi. Tweet yang diajukan kepada LionAir didominasi oleh tweet bersentimen negatif. Proporsi tweet positif dan negatif untuk AirAsia cenderung seimbang. Berdasarkan analisis kompetitif, pandangan publik di Twitter kepada AirAsia cenderung lebih baik dibanding LionAir. Hal tersebut disebabkan AirAsia lebih sering mengadakan promo dan kuis berhadiah. Penelitian ini juga menunjukkan bahwa kernel RBF lebih baik dibanding linier untuk kedua kasus. Metode SMOTE juga mampu meningkatkan nilai AUC dalam kasus imbalanced data. Melalui hasil klasifikasi pada penelitian ini, AirAsia dan LionAir diharapkan mampu mendapatkan informasi mengenai pendapat publik di Twitter secara otomatis sehingga dapat melakukan tindakan perbaikan maupun peningkatan kualitas pelayanan dengan cepat.
=================================================================================================================================
Along with the rapid development of technology and online media, public comments and complaints can be monitored from social media, including Twitter which is often used as a communication media in the business field. AirAsia and LionAir are two airlines operating in Indonesia whose one of their communication media is Twitter. A text mining approach was applied in this study, where the sentiment tweets will be classified using Support Vector Machine with SMOTE (Synthetic Mino-rity Over-sampling Technique) to handle imbalanced data. N-gram con-cept is applied as well in the pre processing step. The classifier are being evaluated by AUC (Area Under Curve). Therefore, the selected classification model is the one gives the highest AUC value. LionAir’s tweets were dominated by negative ones. The proportion of positive and negative tweets for AirAsia tends to be balanced. Based on competitive analysis, public views to AirAsia on Twitter tend to be better than LionAir due to promos and quizzes with prices held by AirAsia. This study shows that the performance of RBF Kernel classification is better than Linear’s for both cases. It is also shown that SMOTE method is able to increase the AUC value for imbalanced data case. Thus, both AirAsia and LionAir are expected to be able to get information of public opinion on Twitter auto-matically so that corrective actions and service quality improvements can be carried out quickly.
| Item Type: | Thesis (Other) |
|---|---|
| Additional Information: | RSSt 006.312 Put t-1 2019 |
| Uncontrolled Keywords: | AirAsia, LionAir, N-gram, SMOTE, Support Vector Machine |
| Subjects: | Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science) |
| Divisions: | Faculty of Mathematics and Science > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Cahya Buana Putri |
| Date Deposited: | 18 Oct 2023 01:38 |
| Last Modified: | 18 Oct 2023 01:38 |
| URI: | http://repository.its.ac.id/id/eprint/64037 |
Actions (login required)
 |
View Item |