Mahdiyyah, Ashila (2026) Pengembangan Sistem Rekomendasi Dessert Dengan Metode Hybrid Content-Based, Demographic-Based, dan Collaborative Filtering Menggunakan Data Ulasan Google Maps di Kawasan Blok M. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5026221148-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5026221148-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy |
Abstract
Dessert merupakan hidangan penutup yang memiliki beragam jenis, rasa, dan tingkat kemanisan. Pilihan dessert yang semakin beragam di kawasan Blok M membuat pengguna kesulitan menemukan tempat yang sesuai dengan preferensi mereka. Penelitian ini mengembangkan sistem rekomendasi dessert dengan pendekatan hybrid yang menggabungkan Content-Based Filtering, Demographic-Based Filtering, dan Collaborative Filtering menggunakan data ulasan Google Maps. Content-Based Filtering memanfaatkan ekstraksi menu dengan Large Language Model (LLM), TF-IDF, One-Hot Encoding, dan Cosine Similarity untuk mengukur kemiripan menu dan rasa. Demographic-Based Filtering memanfaatkan data longitude, latitude, dan Google Maps URL untuk mempertimbangkan kedekatan lokasi, sedangkan Collaborative Filtering memanfaatkan data rating pengguna pada setiap tempat dessert. Ketiga metode digabungkan menggunakan skema hybrid untuk menghasilkan rekomendasi akhir. Sistem dievaluasi menggunakan System Coverage, Diversity (Intra-List Similarity), Catalog Coverage, dan System Usability Scale (SUS). Berdasarkan hasil evaluasi terhadap 332 skenario pengujian, sistem memperoleh System Coverage sebesar 98%, Diversity sebesar 0,3931, Catalog Coverage sebesar 95,70%, dan skor SUS sebesar 85,875 (kategori “Excellent”) melibatkan 40 responden. Hasil tersebut menunjukkan sistem mampu menghasilkan rekomendasi yang relevan, sesuai preferensi pengguna, memanfaatkan sebagian besar item katalog, serta mudah digunakan, sehingga dapat membantu pengguna menemukan dessert yang sesuai preferensinya di kawasan Blok M.
=====================================================================================================================================
Dessert is a dish served after a main course, available in various types, flavors, and levels of sweetness. The growing variety of dessert options in the Blok M area makes it difficult for users to find places that match their preferences. This research develops a dessert recommendation system using a hybrid approach that combines Content-Based Filtering, Demographic-Based Filtering, and Collaborative Filtering using Google Maps review data. Content-Based Filtering utilizes menu extraction with a Large Language Model (LLM), TF-IDF, One-Hot Encoding, and Cosine Similarity to measure menu and flavor similarity. Demographic-Based Filtering utilizes longitude, latitude, and Google Maps URL data to consider location proximity, while Collaborative Filtering utilizes user rating data for each dessert place. The three methods are combined using a hybrid scheme to produce the final recommendations. The system was evaluated using System Coverage, Diversity (Intra-List Similarity), Catalog Coverage, and the System Usability Scale (SUS). Based on the evaluation results of 332 test scenarios, the system achieved a System Coverage of 98%, a Diversity of 0.3931, a Catalog Coverage of 95.70%, and an SUS score of 85,875 which falls into the “Excellent” category based on responses from 40 participant. These results show that the system is able to produce relevant recommendations that align with user preferences, utilize most of the catalog items, and are easy to use, thereby helping users find desserts that suit their preferences in the Blok M area.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Sistem Rekomendasi Dessert, Google Maps, Hybrid Filtering, Content-Based Filtering, Demographic-Based Filtering, Collaborative Filtering, System Usability Scale, Dessert Recommendation System |
| Subjects: | Q Science > Q Science (General) Q Science > QA Mathematics Q Science > QA Mathematics > QA76.9.I58 Recommender systems (Information filtering) |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 57201-(S1) Undergraduate Thesis |
| Depositing User: | Ashila Mahdiyyah |
| Date Deposited: | 28 Jul 2026 03:13 |
| Last Modified: | 28 Jul 2026 03:13 |
| URI: | http://repository.its.ac.id/id/eprint/138336 |
Actions (login required)
 |
View Item |