Article Per Year (5 Year)
p-Index From 2021 - 2026
1.756
P-Index
This Author published in this journals
All Journal ESENSI: JURNAL BISNIS DAN MANAJEMEN Jurnal Informatika dan Teknik Elektro Terapan SMATIKA AGRIKAN Jurnal Ilmiah Agribisnis dan Perikanan Media Mahardhika KOMPUTEK Jurnal Informatika dan Rekayasa Perangkat Lunak Jurnal Teknika BERDAYA: Jurnal Pendidikan dan Pengabdian Kepada Masyarakat Jurnal Scientia Joutica : Journal of Informatic Unisla Journal of Artificial Intelligence and Engineering Applications (JAIEA) Jurnal Cendekia Ilmiah Smatika Jurnal : STIKI Informatika Jurnal
Kemal Farouq Mauladi
Magister Manajemen Universitas Islam Lamongan
Agriculture, Biological Sciences & Forestry Humanities Automotive Engineering Civil Engineering, Building, Construction & Architecture Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Economics, Econometrics & Finance Education Electrical & Electronics Engineering Mathematics Mechanical Engineering Public Health Social Sciences

Published : 21 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 21 Documents
Search

 1   2   3 
PENINGKATAN KINERJA METODE LEXICON-BASED DENGAN KOREKSI EJAAN JARO-WINKLER PADA ANALISIS SENTIMEN MEDIA SOSIAL Virgiawan Asegaf; Kemal Farouq Mauladi; Nur Qomariyah Nawafilah
Jurnal Informatika dan Teknik Elektro Terapan Vol. 14 No. 3 (2026)
Publisher : Universitas Lampung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.23960/jitet.v14i3.10585

Abstract

Opini publik di media sosial mengenai isu sensitif seperti dugaan ijazah palsu tokoh publik memiliki dampak besar terhadap dinamika kepercayaan masyarakat. Namun, analisis sentimen berbasis kamus pada teks komentar Instagram kerap terhambat oleh tingginya ketidakteraturan bahasa dan kesalahan pengetikan (typo), yang menurunkan sensitivitas deteksi sistem konvensional. Penelitian ini menerapkan pendekatan analisis sentimen berbasis leksikon yang dioptimasi dengan algoritma jarak string Jaro-Winkler untuk mengevaluasi 5.485 komentar bersih dari akun Instagram @tvonenews. Hasil komputasi menunjukkan bahwa integrasi Jaro-Winkler dengan batas panjang kata ≥ 4 karakter dan ambang batas kemiripan ≥ 0,90 berhasil mendeteksi dan menyelamatkan 570 komentar bermuatan opini yang sebelumnya keliru diklasifikasikan sebagai opini netral oleh leksikon murni. Klasifikasi akhir menghasilkan distribusi sentimen yang didominasi oleh kelas Negatif sebesar 43,8% (2.405 komentar), disusul oleh sentimen Netral sebesar 36,9% (2.025 komentar), dan Positif sebesar 19,2% (1.055 komentar). Penelitian ini membuktikan bahwa kombinasi Leksikon dan Jaro-Winkler sangat efisien meningkatkan kinerja deteksi pada kosakata tidak baku tanpa memerlukan proses pelatihan model data, sekaligus merepresentasikan tingginya kritik dan tuntutan masyarakat akan transparansi hukum. Public opinion on social media regarding sensitive issues, such as alleged fake diplomas of public figures, significantly impacts the dynamics of societal trust. However, dictionary-based sentiment analysis on Instagram comments is frequently impeded by high linguistic irregularity and typographical errors (typos), which degrade the detection sensitivity of conventional systems. This study applies a lexicon-based sentiment analysis approach optimized with the Jaro-Winkler string distance algorithm to evaluate 5,485 clean comments from the @tvonenews Instagram account. The computational results demonstrate that the Jaro-Winkler integration with a word length limit of ≥ 4 characters and a similarity threshold of ≥ 0.90 successfully detected and recovered 570 opinionated comments that were previously misclassified as neutral by the pure lexicon method. The final classification yielded a distribution dominated by Negative sentiment at 43.8% (2,405 comments), followed by Neutral at 36.9% (2,025 comments), and Positive at 19.2% (1,055 comments). This research proves that combining Lexicon and Jaro-Winkler efficiently enhances detection performance on non-standard vocabulary without requiring model training data, while also highlighting the public's strong criticism and demand for legal transparency.
Co-Authors Abdul Ghofur Affan Bachri Ahmad Maulana Herlianto andi andi Rudianto Angky Soedrijanto Arief Budi Laksono Arinil Haqqoh Chintyadhevi Aulia Rahmi Dina Nurul Afifah Endah Sih Prihatini Faisol Mas’ud Imam Tresno Edy Irawan, Mohamad Rizal Nur Laila Nurhayati M Hasan Wahyudi M. Hasan Wahyudi M. Rizal Nur Irawan Masruroh Masruroh Mohammad Yaskun Nawafilah, Nur Qomariyah Nurul Firdaus Nurul fuad Purnomo Hadi Susilo, Purnomo Hadi Ratna Eka Sari Silvia Karvika Putri Syamsudin Hidayatullah Virgiawan Asegaf Yuhronur Efendi Yuhronur Efendi Zulkifli Lubis