Predictive churn analysis with machine learning methods

Gunay, Melike; Ensari, Tolga

doi:10.1109/siu.2018.8404467

Cited by 3 publications

(4 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The success rate achieved when using this algorithm alone in the Predictive Churn Analysis research is higher compared to the hybrid method developed using logistic regression and Naive Bayes. On the other hand, the artificial neural network approach has the highest prediction accuracy rate (91%) [5].…”

Section: Related Workmentioning

confidence: 99%

Prediction Analysis of Laboratory Equipment Depreciation Using Supervised Learning Methods

Farell,

Jalinus,

Yulastri

et al. 2023

TEM Journal

View full text Add to dashboard Cite

Asset management in Indonesia still poses problems in terms of securing state-owned property. These concerns make it difficult for analysts to predict laboratory equipment depreciation. Therefore, this research aims to create a new model to address this issue. Additionally, to support laboratory managers in gaining insights, a technology-based framework in the form of a laboratory equipment depreciation prediction model has been developed. A new model has been created in this research, which integrates supervised learning models with linear regression algorithms, and subsequently employs a waterfall system development approach. The testing results of the model for predicting laboratory equipment depreciation showed a high level of accuracy, reaching 93%. Furthermore, the comparison between the prediction model and the laboratory equipment data tested directly by technicians demonstrated an accuracy rate of 100%. Finally, the numerical results demonstrate that our framework provides a valuable solution to the difficulties in predicting laboratory equipment depreciation, offering an innovative and practical approach to laboratory equipment maintenance.

show abstract

Section: Related Workmentioning

confidence: 99%

Prediction Analysis of Laboratory Equipment Depreciation Using Supervised Learning Methods

Farell,

Jalinus,

Yulastri

et al. 2023

TEM Journal

View full text Add to dashboard Cite

show abstract

“…The effective decision and predictions are taken; advanced data mining techniques are used. An effective heart disease prediction system is designed using neural network which detects and categorizes the risk level of the heart disease [6]. A critical review on various feature selection, feature extraction methods, classification methods and the performances parameters are examined for predicting the wine quality [7]- [13].…”

Section: A Literature Reviewmentioning

confidence: 99%

Heart Disease Prediction and Performance Assessment through Attribute Element Diminution using Machine Learning

Devi¹,

Goud²,

Teja³

et al. 2019

IJITEE

View full text Add to dashboard Cite

In today’s modern world, the human beings are affected with heart disease irrespective of the age. With the advancement of technological growth, predicting the availability of Heart diseases still remains a challenging issue. The difficulty of predicting the heart disease prevails due to the lack of availability of the symptoms. According to World Health Organization, 33% of population died due to heart diseases. For this, the diagnosis of heart diseases is made by complex combination of clinical data. With this overview, we have used Heart Disease Prediction dataset extracted from UCI Machine Learning Repository for predicting the level of heart disease. The prediction of heart disease classes are achieved in four ways. Firstly, the data set is preprocessed with Feature Scaling and Missing Values. Secondly, the raw data set is fitted to classifiers like logistic regression, KNN classifier, Support Vector Machine, Kernel Support Vector Machine, Naive Bayes, Random Forest and Decision Tree classifiers. Third, the raw data set is subjected to dimensionality reduction using Principal Component Analysis to project the dataset with important components. The dimensionality PCA reduced data set is fitted to the above-mentioned classifiers. Fourth, the performance comparison of raw data set and PCA reduced data set is done by analyzing the performance metrics like Precision, Recall, Accuracy and F-score. The implementation is done using python language under Spyder platform with Anaconda Navigator. Experimental results shows that Random forest is found to be effective with the accuracy of 89% without applying PCA, 85% with five component PCA and 86% with seven component PCA.

show abstract

“…Bunlar arasında Naïve Bayes, Karar Ağaçları, Rastgele Orman, Yapay Sinir Ağları, Lojistik Regresyon, Destek Vektör Makineleri ve K-En Yakın Komşuluk en yaygın kullanılan sınıflandırıcılar olarak sayılabilir. Telekomünikasyon sektöründe müşteri kaybı tahmini için son yıllarda yapılan çalışmalar arasında Yapay Sinir Ağları yönteminin tahmin başarısının diğer yöntemlere göre daha yüksek olduğunu görülmektedir [3][4][5]. Ayrıca Günay ve Ensarı, Lojistik Regresyon ve Naïve Bayes tabanlı geliştirilen hibrit bir yöntemin tek başlarına Lojistik Regresyon ve Naïve Bayes yöntemlerine göre daha iyi sonuç vermesine rağmen Yapay Sinir Ağlarından daha iyi sonuç alınamadığı bildirmektedir [5].…”

Section: Gi̇ri̇ş (Introduction)unclassified

“…Telekomünikasyon sektöründe müşteri kaybı tahmini için son yıllarda yapılan çalışmalar arasında Yapay Sinir Ağları yönteminin tahmin başarısının diğer yöntemlere göre daha yüksek olduğunu görülmektedir [3][4][5]. Ayrıca Günay ve Ensarı, Lojistik Regresyon ve Naïve Bayes tabanlı geliştirilen hibrit bir yöntemin tek başlarına Lojistik Regresyon ve Naïve Bayes yöntemlerine göre daha iyi sonuç vermesine rağmen Yapay Sinir Ağlarından daha iyi sonuç alınamadığı bildirmektedir [5]. Yıldız ve Albayrak, Aykırı Değer Analizi ve öznitelik seçiminin sınıflandırma performansına etkisini gözlemlemişlerdir [6].…”

Section: Gi̇ri̇ş (Introduction)unclassified

Class Imbalance Problem In Churn Prediction

Aydin

2022

Politeknik Dergisi

View full text Add to dashboard Cite

Önemli noktalar (Highlights)  Telekomünikasyon sektöründe Müşteri Kaybı Tahmini problemi ele alınmıştır. / We consider Churn Prediction in Telecommunication Industry.  Bu problem verisinde tipik olan Sınıf Dengesizliği Problemine değinilmiştir. / We refer the Class Imbalance Problem which is typical in data of this problem.  Yeniden Örnekleme tekniklerinin sınıf dengesizliği problemini çözmekteki etkisi gözlenmiştir. / We observe the effect of Resampling Techniques on handling class imbalance.  Makine Öğrenmesi tekniklerinin karşılaştırmalı çalışması sunulmaktadır. / This is a comparative study of Machine Learning Techniques.  Yeniden Örnekleme ve Çapraz Geçerlemenin birlikte doğru kullanımına değinilmiştir. / We refer to correct way of using Resampling and Cross Validation jointly.

show abstract

Predictive churn analysis with machine learning methods

Cited by 3 publications

References 1 publication

Prediction Analysis of Laboratory Equipment Depreciation Using Supervised Learning Methods

Prediction Analysis of Laboratory Equipment Depreciation Using Supervised Learning Methods

Heart Disease Prediction and Performance Assessment through Attribute Element Diminution using Machine Learning

Class Imbalance Problem In Churn Prediction

Contact Info

Product

Resources

About