Dalam Era digital, PT PLN (Persero) mengembangkan aplikasi PLN Mobile sebagai layanan online yang menyediakan fitur pembayaran tagihan listrik, pengaduan gangguan, dan pemantauan penggunaan listrik. Penelitian ini bertujuan untuk menganalisis sentimen ulasan pengguna aplikasi PLN Mobile dengan menambahkan kombinasi spelling corrector Peter Norvig dan N-Gram ke dalam proses preprocessing, menggunakan pelabelan SentiStrength_id. Data yang digunakan berjumlah 12.354 ulasan aplikasi PLN Mobile, yang dikumpulkan melalui scraping di Google Play Store dari bulan Januari hingga bulan Juni 2024. Tahapan penelitian meliputi preprocessing, mencakup case folding, removal punctuation, normalisasi, spelling corrector, tokenizing, stopword removal, detokenized, dengan pelabelan SentiStrength_id. Pemodelan dilakukan dengan algoritma Naïve Bayes. Akurasi yang didapatkan sebesar 81%, precission 69%, recall 66% dan f1-Score 66%. Hasil analisis didapatkan label positif sebesar 68,86%, netral sebesar 22,33%, dan label negatif sebesar 8,81%, dan ditemukan bahwa persentase ketidaksesuaian pengguna yang memberikan rating yang tidak sejalan dengan ulasannya adalah 17,54% ulasan positif, 1,79% ulasan negatif, dan 19,33% ulasan netral. PLN Mobile perlu meningkatkan kinerja aplikasi dengan meningkatkan pengalaman pengguna, memperbaiki aplikasi terutama fitur swacam, mempercepat waktu respon, menyederhanakan proses pengaduan, dan meningkatkan komunikasi terkait gangguan listrik serta kualitas pelayanan dari petugas kepada pelanggan.

In the digital era, PT PLN (Persero) developed the PLN Mobile application as an online service that provides features for paying electricity bills, complaining about disturbances, and monitoring electricity usage. This research aims to analyze the sentiment of user reviews of the PLN Mobile application by adding a combination of Peter Norvig's spelling corrector and N-Gram to the preprocessing process, using SentiStrength_id labeling. The data used amounted to 12,354 reviews of the PLN Mobile application, which were collected through scraping on the Google Play Store from January to June 2024. The research stages include preprocessing, including case folding, punctuation removal, normalization, spelling corrector, tokenizing, stopword removal, detokenized, with SentiStrength_id labeling. Modeling is done with Naïve Bayes algorithm. The accuracy obtained is 81%, precission 69%, recall 66% and f1-Score 66%. The results of the analysis obtained a positive label of 68.86%, neutral of 22.33%, and a negative label of 8.81%, and it was found that the percentage of mismatches of users who gave ratings that were not in line with their reviews was 17.54% positive reviews, 1.79% negative reviews, and 19.33% neutral reviews. PLN Mobile needs to improve application performance by improving user experience, improving applications, especially self-service features, speeding up response times, simplifying the complaint process, and improving communication related to electricity disruptions and the quality of service from officers to customers.