Munir, Muhammad Mishbahul (2018) Implementasi Metode Backpropagation Neural Network Berbasis Lexicon Based Features Dan Bag Of Words Untuk Identifikasi Ujaran Kebencian Pada Twitter. Sarjana thesis, Universitas Brawijaya.
Abstract
Ujaran kebencian adalah bahasa yang mengekspresikan suatu kebencian terhadap suatu kelompok atau individu yang bermaksud untuk menghina atau mempermalukan dan medianya bisa terdapat dimana saja, salah satunya Twitter. Twitter merupakan media sosial yang memungkinkan pengguna untuk mengutarakan perasaan dan opini melalui tweet, termasuk tweet yang mengandung ujaran kebencian. Data dokumen atau tweet berasal dari penelitian yang terdahulu tentang ujaran kebencian. Metode yang digunakan dalam mengolah data dokumen tersebut adalah Backpropagation Neural Network dengan pembaruan fitur menggunakan Lexicon Based Features yang dikombinasikan dengan Bag of Words. Pada penelitian ini menggunakan data sebanyak 500 data yang dibagi menjadi data latih sebanyak 400 data dan data uji sebanyak 100 data. Dari hasil pengujian evaluasi, ketika menggunakan Lexicon Based Features nilai rata-rata f-measure sebesar 0%, lebih buruk dibandingkan dengan menggunakan Bag of Words yang nilai rata-rata f-measure sebesar 76,638%, sedangkan ketika Lexicon Based Features dikombinasikan dengan Bag of Words mendapat nilai rata-rata terbaik diantara fitur sebelumnya dengan f-measure sebesar 78,081%. Dan hasil perbandingan metode Backpropagation Neural Network berbasis Lexicon Based Features dan Bag of Words tidak lebih baik dibandingkan dengan Random Forest Decision Tree menggunakan n-gram fitur pada penelitian sebelumnya.
English Abstract
Hate speech is a language that expresses a hatred of a group or individual who intends to insult or humiliate and the media can be found anywhere, one of them Twitter. Twitter is a social media that allows users to express feelings and opinions through tweets, including tweets that contain hate speech. Document or tweet data comes from previous research on hate speech. The method used in processing the document data is Backpropagation Neural Network with feature updates using Lexicon Based Features combined with Bag of Words. In this study using data as much as 500 data is divided into training data as much as 400 data and test data as much as 100 data. From the evaluation test results, when using Lexicon Based Features, the average value of f-measure is 0%, worse than using the Bag of Words with an average f-measure of 76.638%, while when Lexicon Based Features is combined with the Bag of Words got the best average score among the previous features with a f-measure of 78.081%. And the result Backpropagation Neural Network using Lexicon Based Features combined with Bag of Words is not better than Random Forest Decision Tree using n-gram from previous research.
| Item Type: | Thesis (Sarjana) |
|---|---|
| Identification Number: | SKR/FTIK/2018/190/051801212 |
| Uncontrolled Keywords: | Ujaran Kebencian, Twitter, Bag Of Words, Lexicon Based Features, Backpropagation Neural Network |
| Subjects: | 000 Computer science, information and general works > 006 Special computer methods > 006.3 Artificial intelligence > 006.32 Neural nets (neural networks) |
| Divisions: | Fakultas Ilmu Komputer > Teknik Informatika |
| Depositing User: | Yusuf Dwi N. |
| Date Deposited: | 31 May 2018 06:35 |
| Last Modified: | 26 Oct 2021 09:36 |
| URI: | http://repository.ub.ac.id/id/eprint/11305 |
Preview |
Text
BAB V.pdf Download (972kB) | Preview |
Preview |
Text
BAB VII.pdf Download (553kB) | Preview |
Preview |
Text
Bagian Depan.pdf Download (1MB) | Preview |
Preview |
Text
Daftar Pustaka.pdf Download (549kB) | Preview |
Preview |
Text
BAB VI.pdf Download (815kB) | Preview |
Preview |
Text
Lampiran.pdf Download (1MB) | Preview |
Preview |
Text
BAB I.pdf Download (584kB) | Preview |
Preview |
Text
BAB II.pdf Download (959kB) | Preview |
Preview |
Text
BAB III.pdf Download (560kB) | Preview |
Preview |
Text
BAB IV.pdf Download (1MB) | Preview |
Actions (login required)
 |
View Item |