Fiber internet müşteri şikayet tahminlemesi
Küçük Resim Yok
Tarih
2021
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
İstanbul Beykent Üniversitesi
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Bu çalışmada telekomünikasyon sektöründe faaliyet gösteren bir firmaya gelen fiber internet müşteri şikayetleri verileri, kişisel verilerin korunması kanunu kapsamında anonimleştirilmiştir. Makine öğrenmesi algoritmalarından karar ağaçları, naive bayes, random forest, lojistik regresyon ve xgboost yöntemlerini python programlama dili ile modelleyerek müşteri şikayetlerinin gerçekten bir problem kaynaklı gelip gelmediğine bakarak tahmin edilmesi sağlanmıştır. Çalışmada öncelikle makine öğrenmesi algoritmalarıyla tekil modeller oluşturarak başarı oranları hesaplanmıştır. Daha sonra bu algoritmalardan birleşik (ikili) hibrit modeller oluşturularak başarı oranları karşılaştırılmıştır. Çalışmanın amacı, müşteri şikayetlerine hızlı müdahale edilmesini sağlamaktır. Ayrıca müşteri memnuniyetini arttırarak müşteri odaklı bir yaklaşım benimsemektir.
In this study, fiber internet customer complaints data received by a company operating in the telecommunications sector were anonymized within the scope of the personal data protection law. By modeling the machine learning algorithms decision trees, naive bayes, random forest, logistic regression and xgboost methods with the python programming language, it is ensured that customer complaints are predicted by looking at whether they really come from a problem. In the study, success rates were calculated by first creating single models with machine learning algorithms. Then, combined (binary) hybrid models were created from these algorithms and their success rates were compared. The aim of the study is to ensure rapid response to customer complaints. In addition, it is to adopt a customer-oriented approach by increasing customer satisfaction.
In this study, fiber internet customer complaints data received by a company operating in the telecommunications sector were anonymized within the scope of the personal data protection law. By modeling the machine learning algorithms decision trees, naive bayes, random forest, logistic regression and xgboost methods with the python programming language, it is ensured that customer complaints are predicted by looking at whether they really come from a problem. In the study, success rates were calculated by first creating single models with machine learning algorithms. Then, combined (binary) hybrid models were created from these algorithms and their success rates were compared. The aim of the study is to ensure rapid response to customer complaints. In addition, it is to adopt a customer-oriented approach by increasing customer satisfaction.
Açıklama
Anahtar Kelimeler
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrol, Computer Engineering and Computer Science and Control
