Analisis Sentimen Media Sosial Opini Ujian Nasional Berbasis Komputer menggunakan Metoda Naive Bayes
Abstract
Ujian Nasional merupakan proses tolak ukur kemampuan hasil belajar siswa dan siswi selama proses belajar di sekolah, seiring perkembangan zaman, terdapat perubahan pada proses ujian nasional, yaitu sistem pengerjaan ujian nasional secara terkomputerisasi atau dapat di sebut dengan Computer Based Test (CBT). Dengan adanya Ujian Nasioal Berbasis komputer tentu menjadi bahasan-bahasan baru bagi masyarakat, baik bahasan pro dan kontra sehingga banyak masyarakat yang melontarkan opini-opininya melalui media sosial. Penelitian ini telah membahas mengenai analisa sentimen opini ujian nasional berbasis komputer. Sample yang di ambil sebanyak 181 kalimat sentimen yang di olah menggunakan algoritma Naive Bayes dengan mengelompokkan data menjadi tiga kelas Sentimen Positif, Netral, Negatif. Hasil pengolahan data menunjukan kelas sentimen netral memiliki nilai tertinggi sebesar 79% dan nilai terendah di peroleh kelas negatif dengan nilai 0.09%. Sedangkan tingkat akurasi ketiga kelas sentimen mencapai 100%.
Keywords— Naive Bayes, Opini, Sentimen, Twitter
The National exam is a benchmark of students’ learning capability result during learning process in schools. As the era develops, national exam is changing in its process. Nowadays, the test uses Computer Based Test (CBT) system. The existence of this system leads to new topics for society. Consequently, pro and contra opinions are thrown by them through social media. Therefore, this research discuses about analyses of sentiment opinions on CBT national exam. Naive Bayes’ Algorithm is used for processing 181 data samples in form of sentiment sentences. The data samples are grouped into three classess; Positive, Neutral, and Negative. The results of data processing have shown that Neutral sentiment gains the highest percentage 79% while the Negative sentiment is the lowest with value 0.09% Overall, accuracy degree of the three sentiment classes reach 100%.
Keywords— Naive Bayes, Opini, Sentimen, Twitter
Full Text:
PDFDOI: http://dx.doi.org/10.33021/jeee.v1i2.189
Refbacks
- There are currently no refbacks.