KLASIFIKASI SENTIMEN SUPPORT VECTOR MACHINE BERBASIS PARTICLE SWARM OPTIMIZATION MENYAMBUT PEMILU 2024

SILALAHI, WILLIAM (2023) KLASIFIKASI SENTIMEN SUPPORT VECTOR MACHINE BERBASIS PARTICLE SWARM OPTIMIZATION MENYAMBUT PEMILU 2024. S1 thesis, Universitas Mercu Buana Bekasi.

Preview	Text 41519210046 - William Silalahi - 01 Cover.pdf Download (337kB) | Preview
Preview	Text 41519210046 - William Silalahi - 02 Abstrak.pdf Download (153kB) | Preview
	Text 41519210046 - William Silalahi - 03 BAB 1.pdf Restricted to Registered users only Download (158kB)
	Text 41519210046 - William Silalahi - 04 BAB 2.pdf Restricted to Registered users only Download (325kB)
	Text 41519210046 - William Silalahi - 05 BAB 3.pdf Restricted to Registered users only Download (381kB)
	Text 41519210046 - William Silalahi - 06 BAB 4.pdf Restricted to Registered users only Download (1MB)
	Text 41519210046 - William Silalahi - 07 BAB 5.pdf Restricted to Registered users only Download (125kB)
	Text 41519210046 - William Silalahi - 08 Daftar Pustaka.pdf Restricted to Registered users only Download (267kB)
	Text 41519210046 - William Silalahi - 09 Lampiran.pdf Restricted to Registered users only Download (895kB)

Abstract

Ajang pesta demokrasi di tahun 2024 yang disertai dengan narasi politik akan kembali diadakan. Beragam komentar negatif dan isu hoaks mulai bermunculan di media sosial untuk menjatuhkan pihak oposisi. Hal ini menimbulkan ketidakpercayaan masyarakat hingga muncul golongan putih. Oleh karena itu, klasifikasi sentimen berdasarkan komentar di Twitter dan kuesioner dilakukan agar mengetahui bagaimana pandangan masyarakat mengenai fenomena ini. Algoritma yang dipakai adalah SVM dan PSO dengan metode tambahan seperti TF-IDF dalam pembuatan vektor dan SMOTE untuk menyeimbangkan data pada setiap kelas. Karena algoritma machine learning tersebut bersifat supervised learning, maka pelabelan otomatis dilakukan menggunakan VADER sebagai data latih. WordCloud sebagai sarana pendukung dalam mempersiapkan pemilu tahun depan juga diterapkan. Hasil pelabelan otomatis mendapatkan 1178 data sentimen positif dan 422 data sentimen negatif. Lalu secara berturut-turut, akurasi SVM tanpa optimasi dengan metode percentage split validation berdasarkan data latih dan data uji 70%:30%, 80%:20%, 90%:10% adalah sebesar 85.27%, 87.08%, dan 86.02%. Sedangkan persentase SVM (PSO) adalah 86.69%, 80.51%, dan 89.41%. SVM (PSO) dengan split 90%:10% mendapatkan hasil akurasi tertinggi dengan persentase 89.41%. WordCloud sentimen positif dan negatif menunjukkan bahwa masyarakat mendukung adanya pemilu 2024. Kata Kunci: SVM, PSO, VADER, WordCloud, Pemilu The democratic party event in 2024 which is accompanied by a political narrative will be held again. Various negative comments and hoax issues began to appear on social media to bring down the opposition. This created distrust in society and led to the emergence of non-voters. Therefore, the classification of sentiments based on comments on Twitter and questionnaires was carried out in order to find out how the public views this phenomenon. The algorithms used are SVM and PSO with additional methods such as TF-IDF in vector creation and SMOTE to balance data in each class. Since the machine learning algorithm is supervised learning, automatic labeling is done using VADER as training data. WordCloud as a supporting tool in preparing for next year's elections is also implemented. The results of automatic labeling get 1178 positive sentiment data and 422 negative sentiment data. Then successively, the accuracy of SVM without optimization on percentage split validation method based on 70%:30%, 80%:20%, 90%:10% train and test data are 85.27%, 87.08% and 86.02%. While the percentage of SVM (PSO) is 86.69%, 80.51% and 89.41%. SVM (PSO) with a split of 90%:10% gets the highest accuracy with a percentage of 89.41%. WordCloud positive and negative sentiments show that the public supports the 2024 election. Keywords: SVM, PSO, VADER, WordCloud, Election.

Actions (login required)

View Item