Rizal, Naufal Aulia (2014) Implementasi metode hybrid JST- SOM pada prediksi churn pelanggan seluler : studi kasus PT. Telekomunikasi seluler. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
5110100076-Undergraduate_Thesis.pdf Download (2MB) | Preview |
Abstract
Literatur marketing menyatakan bahwa lebih mahal biaya yang dikeluarkan provider untuk mendapatkan pelanggan baru daripada mempertahankan pelanggan loyal yang sudah ada. Churn pelanggan adalah istilah yang digunakan pada industri telekomunikasi yang artinya adalah perpindahan pelanggan dari satu provider ke provider lainnya. Model prediksi churn pelanggan dikembangkan untuk mempertahankan pelanggan yang sudah ada. Manajemen churn sangat penting bagi aktivitas provider untuk mempertahankan pelanggan yang loyal. Sangat dibutuhkan metode untuk memprediksi churn dengan baik. Pasar industri telekomunikasi berkembang sangat kompetitif sehingga membuat manajemen churn pelanggan sangat krusial bagi provider telekomunikasi.
Pada Tugas Akhir ini akan diterapkan metode hybrid data mining yaitu penggabungan dua metode classification dan clustering. Jaringan Saraf Tiruan (JST) akan digabungkan dengan Self-organizing Map (SOM) untuk menyelesaikan permasalahan prediksi churn pelanggan. SOM akan digunakan sebagai metode clustering untuk mengurangi data yang tidak merepresentasikan data latih, lalu output SOM akan dipakai sebagai data latih prediksi churn pelanggan menggunakan propagasi balik JST.Penggabungan metode JST dan SOM pada data uji menghasilkan akurasi hingga 96.7%. Berdasarkan hasil uji coba yang dilakukan penggabungan JST-SOM dapat memaksimalkan akurasi klasifikasi.
===================================================================================================
Marketing literature states that it is more costly to engage a new customer than to retain an existing loyal customer. Customer churn is a term used in the telecommunications industry which means customer migration from one provider to another provider. Churn prediction models are developed by academics and practitioners to effectively manage and control customer churn in order to retain existing customer. As churn management is an important activity for companies to retain loyal customer, the abilty to correctly predict customer churn is necessary. As the cellular network services market becoming more competitive, customer churn management has become a crucial task for mobile communication operators
In this final project will apply a hybrid data mining method that combine two method, classification and clustering. Neural Network (NN) combined with Self-organing Map (SOM) will be applied to solve prediction problem for customer churn. SOM will be used for clustering techniques to perform data reduction task by filtering out unrepresentative training data, then the outputs from SOM are used to create the prediction model based on Backpropagation Neural Network. Hybrid NN and SOM on the test data will generate up to 96.7% accuracy. Based on the test results, the hybrid NN-SOM can maximize classification accuracy.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Additional Information: | RSIf 005.74 Riz i |
| Uncontrolled Keywords: | Hybrid data mining; self-organizing map; jaringan saraf tiruan; churn pelanggan |
| Subjects: | H Social Sciences > HD Industries. Land use. Labor > HD30.335 Communication systems |
| Divisions: | Faculty of Information Technology > Computer Engineering > 90243-(S1) Undergraduate Thesis |
| Depositing User: | - Taufiq Rahmanu |
| Date Deposited: | 26 Jul 2019 02:34 |
| Last Modified: | 26 Jul 2019 02:34 |
| URI: | http://repository.its.ac.id/id/eprint/66069 |
Actions (login required)
 |
View Item |