Monitoring Sentimen Ulasan Pengguna Terhadap Super App Polri di Google Play dengan Diagram Kendali Laney p’ Berbasis Convolutional Neural Network-Bidirectional Long Short-Term Memori (CNN-BiLSTM)

Zalukhu, Hollyviar Resnias Putri (2026) Monitoring Sentimen Ulasan Pengguna Terhadap Super App Polri di Google Play dengan Diagram Kendali Laney p’ Berbasis Convolutional Neural Network-Bidirectional Long Short-Term Memori (CNN-BiLSTM). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003221044_Undergraduate_Thesis.pdf Restricted to Repository staff only Download (7MB) | Request a copy

Abstract

Transformasi digital di sektor pelayanan publik mendorong Kepolisian Republik Indonesia (POLRI) menghadirkan Super App Polri sebagai platform layanan kepolisian berbasis daring. Aplikasi ini menyediakan layanan seperti pengurusan SKCK, SIM, dan STNK, yang memungkinkan interaksi langsung antara pengguna dan sistem digital. Namun, implementasinya memunculkan beragam pengalaman pengguna yang tercermin dalam ulasan di Google Play. Penelitian ini bertujuan menganalisis sentimen ulasan pengguna sekaligus memantau stabilitas kualitas layanan Super App Polri secara berkelanjutan. Metode yang digunakan mengombinasikan pendekatan deep learning melalui model Convolutional Neural Network-Bidirectional Long Short-Term Memory (CNN-BiLSTM) untuk klasifikasi sentimen, serta Statistical Process Control (SPC) menggunakan diagram kendali Laney p’ untuk evaluasi kinerja layanan dari waktu ke waktu. Hasil penelitian menunjukkan bahwa model CNN-BiLSTM berbasis FastText menghasilkan performa dengan kategori Excellent Classification dengan akurasi 0,9228, AUC 0,9608, serta recall kelas negatif sebesar 0,9474, yang menegaskan keandalannya dalam mendeteksi keluhan pengguna. Analisis SPC menunjukkan bahwa proses layanan belum terkendali secara statistik, khususnya saat dievaluasi menggunakan diagram Laney p’ berbasis target mutu layanan. Sebanyak 65 dari 109 minggu berada dalam kondisi out-of-control, didukung oleh analisis kapabilitas proses dengan perolehan nilai Pp=0,24 dan Ppk= -0,84, yang menunjukkan variabilitas proses yang masih lebar atau kurang presisi dan kinerja aktual menyimpang jauh dari standar mutu pelayanan kategori “Baik”. Identifikasi lebih lanjut mengungkap dominasi keluhan terjadi di kategori critical dan marginal, yang berkaitan dengan kegagalan teknis sistem dan inefisiensi prosedural, sehingga menjadi prioritas utama perbaikan kualitas layanan
==================================================================================================================================
Digital transformation in the public service sector has driven the Indonesian National Police (POLRI) to introduce the Super App Polri as an online-based policing service platform. This application provides services such as the issuance of police clearance certificates (SKCK), driving licenses (SIM), and vehicle registration documents (STNK), enabling direct interaction between users and the digital system. However, its implementation has generated diverse user experiences, as reflected in reviews on Google Play. This study aims to analyze user review sentiment while continuously monitoring the stability of service quality of the Super App Polri. The proposed method combines a deep learning approach using a Convolutional Neural Network–Bidirectional Long Short-Term Memory (CNN-BiLSTM) model for sentiment classification with Statistical Process Control (SPC) employing the Laney p’ control chart to evaluate service performance over time. The results indicate that the FastText-based CNN-BiLSTM model achieves Excellent Classification performance, with an accuracy of 0,9228, an AUC of 0,9608, and a negative-class recall of 0,9474, confirming its reliability in detecting user complaints. SPC analysis shows that the service process is not statistically in control, particularly when evaluated using the Laney p’ control chart based on service quality targets. A total of 65 out of 109 weeks are identified as out-of-control, supported by process capability analysis yielding Pp=0,24 and Ppk= -0,84. These results indicate wide process variability (low precision) and that the actual service performance deviates substantially from the “Good” service quality standard. Further analysis reveals that user complaints are predominantly classified into critical and marginal categories, associated with technical system failures and procedural inefficiencies, making them the primary priorities for service quality improvement.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Analisis Sentimen, Analisis Kinirja Proses, CNN-BiLSTM, Diagram Kendali Laney p’, Super App Polri, CNN-BiLSTM, Laney p’ Control Chart, Process Performance Evaluation, Super App Polri, Sentiment Analysis
Subjects:	Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Hollyviar Resnias Putri Zalukhu
Date Deposited:	27 Jan 2026 03:06
Last Modified:	27 Jan 2026 03:06
URI:	http://repository.its.ac.id/id/eprint/129835

Actions (login required)

View Item