Implementasi Text Mining pada Data Perhotelan Menggunakan Support Vector Machine (SVM) dan Analisis Topik dengan Model Probabilistik

Panjaitan, Yana Rezki Kriswin (2021) Implementasi Text Mining pada Data Perhotelan Menggunakan Support Vector Machine (SVM) dan Analisis Topik dengan Model Probabilistik. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 06211740000043-Undergraduate_Thesis.pdf] Text
06211740000043-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2023.

Download (2MB) | Request a copy

Abstract

Kota Bogor merupakan kota yang berpotensi menjadi objek wisata. Covid-19 menyebar secara masif keseluruh dunia termasuk Indonesia sehingga pemerintah mengambil kebijakan seluruh kegiatan dilakukan dari rumah. Perubahan tatanan kehidupan tersebut memberikan tantangan baru bagi pihak hotel untuk tidak memakai cara konvensional seperti kuesioner dalam mengetahui kepuasan tamu sehingga memaksa pihak hotel menggunakan media yang ada salah satunya situs TripAdvisor. Situs ini memuat harga, tipe hotel, ulasan pengunjung dan sebagainya. Hal ini membuat tamu tidak kesulitan mencari informasi hotel, sedangkan pihak hotel diuntungkan dengan ulasan yang ada. Namun jumlah ulasan yang banyak menyita waktu untuk memahami satu persatu ulasan, sehingga diperlukan text mining. Dalam proses text mining ulasan diklasifikasikan terlebih dahulu menjadi sentimen positif dan negatif menggunakan metode SVM. Klasifikasi hanya memberi informasi sentimen positif atau negatif, sehingga dibutuhkan LDA dan LSA untuk menemukan informasi tersembunyi guna mengetahui kepuasan tamu pada pelayanan hotel. Model klasifikasi terbaik dalam penelitian ini menggunakan SVM kernel linear pada data yang telah diatasi imbalancednya dengan SMOTE. Metode LDA menghasilkan topic coherence lebih tinggi dibanding LSA sehingga membentuk topik positif pada Novotel adalah breakfast dan dinner, hotel dapat menjadi tempat rapat, makanan variatif dan hotel ramah anak. Sedangkan topik negatifnya proses high season lambat, makanan habis tidak langsung di refill, hotel tidak sesuai dengan informasi booking online, area kolam pria dan wanita tidak dibedakan, akses menuju tempat bermain anak, drainage tidak baik, paving tidak rata, bau pesing, kamar deluxe kurang baik, bathup teras duduk mengkhawatirkan. Topik positif Grand Savero adalah sarapan variatif, pelayanan cepat, menu enak, ramah anak, konsep indoor modern, booking mudah, terdapat aneka bubur dan buah, kinerja team bagus, fasilitas lengkap, dan ruangan yang segar dan bersih, sedangkan topik negatifnya proses reserve dan akses menuju area hotel sulit dan extra bed berbayar. Topik positif pada Aston adalah menu breakfast variatif dan enak serta hotel ramah anak, sedangkan topik negatifnya sikap staf, menu dinner tidak sesuai harga, fying fox berbayar, area balkon bau, suara mengganggu, request tidak sesuai, tempat bermain anak rusak, bising, perabotan kamar kurang.
=====================================================================================================
Bogor City is a city that has the potential to become a tourist attraction. Covid-19 has spread massively throughout the world, including Indonesia, so the government has made a policy that all activities are carried out from home. The change in the order of life provides a new challenge for the hotel not to use conventional methods such as questionnaires to determine guest satisfaction, thus forcing the hotel to use media, one of which is the TripAdvisor site. This site contains prices, hotel types, visitor reviews and so on. This makes it easy for guests to find hotel information, while the hotel benefits from existing reviews. However, the large number of reviews takes time to understand one by one review, so text mining is needed. In the text mining process, reviews are classified first into positive and negative sentiments using the SVM method. Classification only provides information on positive or negative sentiments, so LDA and LSA are needed to find hidden information to find out guest satisfaction with hotel services. The best classification model in this study uses a linear kernel SVM on data that has been overcome with SMOTE imbalancednya. The LDA method produces higher topic coherence than LSA so that it forms positive topics at Novotel, namely breakfast and dinner, hotels can be used as meeting places, varied food and child-friendly hotels. While the negative topics are the slow high season process, food runs out not immediately refilled, hotels do not match online booking information, male and female pool areas are not distinguished, access to children's playgrounds, drainage is not good, paving is uneven, urine smells, deluxe rooms not good, the sitting terrace bathtub is worrying. The positive topics of Grand Savero are varied breakfasts, fast service, delicious menus, child friendly, modern indoor concepts, easy booking, there are various porridge and fruit, good team performance, complete facilities, and fresh and clean rooms, while the negative topics are the reserve process and access to the hotel area is difficult and extra beds are paid. Positive topics at Aston are varied and delicious breakfast menus and child-friendly hotels, while the negative topics are staff attitudes, dinner menus don't match the price, paid fying fox, smelly balcony area, annoying sound, inappropriate requests, damaged children's play area, noise, furniture less room.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: LDA, LSA, SMOTE, Ulasan, SVM LDA, LSA, Reviews, SMOTE, SVM
Subjects: Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
T Technology > T Technology (General) > T57.8 Nonlinear programming. Support vector machine. Wavelets. Hidden Markov models.
Divisions: Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User: Yana Rezki Kriswin Panjaitan
Date Deposited: 08 Sep 2021 05:37
Last Modified: 08 Sep 2021 05:37
URI: http://repository.its.ac.id/id/eprint/91857

Actions (login required)

View Item View Item