Utami, Shindy Sari (2019) Analisis Sentimen Pengguna Twitter Mengenai “Sedotan Plastik” Dengan Metode K-Nearest Neighbor (KNN) Dan Neighbor-Weighted K-Nearest Neighbor (NWKNN). Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
06211540000005-Undergraduate_Theses.pdf Download (2MB) | Preview |
Abstract
Tahun 2017 Divers Clean Action, kelompok pemerhati lingkungan khususnya laut menyebutkan bahwa pemakaian sedotan di Indonesia mencapai 93.244.847 batang setiap harinya. Beberapa waktu terakhir, terdapat gerakan yang dibuat untuk menggugah kepedulian masyarakat akan lingkungan, yaitu gerakan untuk tidak menggunakan sedotan plastik. Gerakan anti sedotan plastik ini telah direspon oleh masyarakata pada media social twitter. Respon masyarakat diklasifikasikan kedalam sentimen positif dan negatif menggunakan metode K-Nearest Neighbor (KNN) serta Neighbor-Weighted K-Nearest Neighbor (NWKNN). Sentimen positif memiliki rasio 81% atau sebesar 2587 tweet, sedangkan sentimen negatif sebesar 19% atau 599 tweet. Metode NWKNN memberikan hasil lebih baik daripada metode KNN ketika nilai K masih bernilai kecil yaitu K=1 sampai K=3. Setelah K=4 hasil AUC metode NWKNN mengalami penurunan lebih banyak dibandingkan dengan metode KNN. Hasil visualisasi wordcloud sentimen positif terdapat kata “besi”, “bambu”, dan “kokop”. Wordcloud sentimen negatif terdapat kata “dingin”, “gigi”, dan “ngilu”.
================================================================================================================================
In 2017 Divers Clean Action, a group of environmentalists, especially the sea, said that the use of straws in Indonesia reached 93,244,847 sticks every day. For the last time, there was a movement that was made to arouse public concern about the environment, namely the movement not to use plastic straws. This action has been responded by the public on twitter social media. Community responses are classified into positive and negative sentiments using the K-Nearest Neighbor (KNN) method and Neighbor-Weighted K-Nearest Neighbor (NWKNN) method. Positive sentiment has a ratio of 81% or 2587 tweets, while the negative sentiment is 19% or 599 tweets. The NWKNN method gives better results than the KNN method when the K value is still small, between K = 1 until K = 3. After K = 4 the results of the AUC NWKNN method decreased more than the KNN method. The results of the visual wordcloud positive sentiments include the words "iron", "bamboo", and "kokop". The negative sentiments wordcloud has the words "cold", "teeth", and "pain".
| Item Type: | Thesis (Other) |
|---|---|
| Additional Information: | RSSt 519.536 Uta a-1 2019 |
| Uncontrolled Keywords: | Sedotan plastik, klasifikasi, KNN, NWKNN, Twitter |
| Subjects: | Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) |
| Divisions: | Faculty of Mathematics, Computation, and Data Science > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Utami Shindy Sari |
| Date Deposited: | 05 Jun 2023 01:06 |
| Last Modified: | 05 Jun 2023 01:06 |
| URI: | http://repository.its.ac.id/id/eprint/64038 |
Actions (login required)
 |
View Item |