CD Skripsi

Klasifikasi Target Layanan Pelanggan Menggunakan Convolutional Neural Network Di Pt. Telkom Indonesia Tbk. Witel Riau

---

IndiHome and IndiBiz faced challenges in customizing service offerings. Inappropriate marketing strategies lead to missed sales targets and the increasing amount of data makes manual monitoring ineffective. In the face of these challenges, advances in Deep Learning technology support the automation of various computer vision tasks, including recognizing digital objects. This research applies Convolutional Neural Network (CNN) model and MobileNetV2 architecture for building image classification. The dataset consists of 882 images categorized into two classes and divided into three parts, 60% for training, 20% for validation, and 20% for testing. Optimization is done based on batch size and learning rate to improve model performance. The best results were obtained at batch size 32 and learning rate 0.0001 on MobileNetV2 architecture. This model shows superior performance, with test results using testing data showing an accuracy value of 92.05%, as well as precision, recall, and f1-score values of 91.46% each. The performance of the model is measured based on the confusion matrix on the testing data. The results show that false positives and false negatives in both classes have low error rates, so they do not significantly affect the overall performance of the model. Thus, the MobileNetV2 architecture is proven to optimally support digital object recognition automation.

Keywords: Convolutional Neural Network, IndiBiz, IndiHome, Internet Service Provider, MobileNetV2.

---

Ketersediaan

Informasi Detail

Judul Seri

-

No. Panggil

2103110233

Penerbit

Pekanbaru : Universitas Riau FMIPA Sistem Informasi., 2025

Deskripsi Fisik

-

Bahasa

Indonesia

ISBN/ISSN

-

Klasifikasi

2103110233

Tipe Isi

-

Tipe Media

-

Tipe Pembawa

-

Edisi

-

Subjek

SISTEM INFORMASI

Info Detail Spesifik

-

Pernyataan Tanggungjawab

Mutia

Versi lain/terkait

Lampiran Berkas

Komentar

---

Anda harus masuk sebelum memberikan komentar