ANALISIS SENTIMEN MASYARAKAT TERHADAP TABUNGAN PERUMAHAN RAKYAT (TAPERA) DENGAN PERBANDINGAN METODE SUPPORT VECTOR MACHINE DAN K-NEAREST NEIGHBORS PADA SOSIAL MEDIA YOUTUBE

Rizqullah Kasmir; Rahmi Imanda;

ANALISIS SENTIMEN MASYARAKAT TERHADAP TABUNGAN
PERUMAHAN RAKYAT (TAPERA) DENGAN PERBANDINGAN
METODE SUPPORT VECTOR MACHINE DAN K-NEAREST
NEIGHBORS PADA SOSIAL MEDIA YOUTUBE

Rizqullah Kasmir

Perumahan merupakan masalah yang dihadapi oleh beberapa negara berkembang termasuk
Indonesia. Untuk mengatasi hal tersebut pemerintah mengeluarkan kebijakan program Tabungan
Perumahan Rakyat (TAPERA) yang ternyata mendapat banyak opini dari berbagai pihak. Tujuan
dari penelitian ini untuk melihat akurasi dari metode Support Vector Machine dan K-Nearst
Neighbors. Didapat 3103 data yang sudah melalui proses pelabelan yang terdiri 2.003 sentimen
positif dan 1100 sentimen negatif. Data tersebut dibagi menjadi dua yaitu data latih dan data testing,
untuk data testing sebanyak 20% dan data latih 80%. Sebelum dilakukan proses kedalam dua metode
tersebut, dilakukan tahap proses pengumpulan data, preprocessing, normalisasi, pelabelan,

menggunakan kamus lexicon. Setelah dilakukan proses klasifikasi Support Vector Machine & K-
Nearst Neighbors, didapati nilai akurasi Support Vector Machine 82%, precision 83% dan recall

92% sementara K-Nearst Neighbors dengan nilai akurasi 77%, precision 78%, dan recall 90%.
Menunjukan bahwa metode Support Vector Machine memiliki tingkat akurasi lebih tinggi dalam
menganalisis sentimen masyarakat terhadap TAPERA.
Kata kunci: TAPERA, SVM, KNN

ANALYSIS OF PUBLIC SENTIMENT TOWARDS PUBLIC HOUSING
SAVINGS (TAPERA) WITH A COMPARISON OF SUPPORT VECTOR
MACHINE AND K-NEAREST NEIGHBORS METHODS ON YOUTUBE

SOCIAL MEDIA
Rizqullah Kasmir

Housing is a problem faced by several developing countries including Indonesia. To overcome this,
the government issued a policy on the Public Housing Savings Program (TAPERA) which
apparently received a lot of opinions from various parties. The purpose of this research is to see the
accuracy of the Support Vector Machine and K-Nearest Neighbors methods. There are 3103 data
that have gone through the labeling process consisting of 2,003 positive sentiments and 1100
negative sentiments. The data is divided into two, namely training data and testing data, for 20%
testing data and 80% training data. Before the process is carried out into the two methods, the data
collection process, preprocessing, normalization, labeling, using a lexicon dictionary. After the
classification process of Support Vector Machine & K-Nearest Neighbors, the accuracy value of
Support Vector Machine is 82%, precision 83% and recall 92% while K-Nearest Neighbors with an
accuracy value of 77%, precision 78%, and recall 90%. Shows that the Support Vector Machine
method has a higher level of accuracy in analyzing public sentiment towards TAPERA.
Keywords: TAPERA, SVM, KNN