Nabila, Herviana Mayu (2020) Movie Genre Multilabel Classification Using Problem Transformation and Multilabel K-Nearest Neighbor (ML-KNN). Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
06211640000075_Undergraduate Thesis.pdf Download (2MB) | Preview |
Abstract
Film merupakan bentuk media massa yang mampu memberikan nilai hiburan pada masyarakat. Berkembangnya dunia digital membuat masyarakat dapat dengan mudah meng-akses situs online yang menyediakan informasi mengenai genre film, salah satunya melalui situs rating film seperti IMDb dan TMDb. Pengelompokkan genre film dilakukan secara manual dan tidak efisien karena membutuhkan banyak waktu dan tenaga ahli, sehingga klasifikasi genre film secara otomatis dapat menjadi solusi. Suatu film tidak hanya memiliki satu genre saja namun dapat memiliki lebih dari satu genre, sehingga klasifikasi genre film dikategorikan sebagai klasifikasi multilabel. Pada penelitian tugas akhir ini dilakukan klasifikasi genre drama, action, adventure, thriller, dan comedy berdasarkan teks sinopsis film dengan membandingkan pendekatan transformasi dan adaptasi algoritma. Metode transformasi yang digunakan adalah Label Powerset (LP), dan metode adaptasi algortima yang digunakan adalah Multilabel K-Nearest Neighbor (ML-KNN). Data yang digunakan berupa teks sinopsis dan genre film yang diambil dari situs The Movie Database (TMDb). Hasil penelitian menunjukkan bahwa metode Multilabel K-Nearest Neughbor (ML-KNN) meng-hasilkan hamming loss lebih kecil dibandingkan metode K-Nearest Neighbor dengan transformasi Label Powerset (LP).
=======================================================
A movie is a media that is able to provide entertaiment value to the public. Movies can be devided into several categories or called movie genres. Movie gere is useful for people to watch movies based on the type they like. The development of digital world allows people to easily access online sites that provide information about movie genres, one which is through a movie rating websites such as IMDb and TMDb. Manual genre classification is ineffiecient because it requires a lot of time, so automatic movie genre classification can be a solution. A movie does not only have one genre but can have several genres at once, so the classification of movie genre is categorized as multilabel classification. In multilabel classification, a data can be categori-zed to more than one label. This study will comparing the problem transformation method called Binary Relevance (BR) and adaptation algorithm Method named Multilabel K-Nearest Neighbor (ML-KNN) on classfying drama, action, adventure, thriller, and comedy movie genre. The data used in this study is synopsis text and movie genres that taken from The Movie Database (TMDb) sites using TMDb API. The study conducted shows that movie genre multilabel classification using Multilabel K-Nearest Neighbor
| Item Type: | Thesis (Other) |
|---|---|
| Additional Information: | RSSt 519.53 Nab k-1 • Nabila, Herviana Mayu |
| Uncontrolled Keywords: | Genre Film, Label Powerset, ML-KNN, Multilabel ================================================= Label Powerset, ML-KNN, Multilabel, Movie Genre |
| Subjects: | Q Science > QA Mathematics > QA279.5 Bayesian statistical decision theory. Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science) |
| Divisions: | Faculty of Mathematics, Computation, and Data Science > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Herviana Mayu Nabila |
| Date Deposited: | 26 Aug 2020 06:33 |
| Last Modified: | 03 Jan 2024 07:12 |
| URI: | http://repository.its.ac.id/id/eprint/81162 |
Actions (login required)
 |
View Item |