MACHINE LEARNING TECHNIQUES FOR CUSTOMER CHURN PREDICTION IN THE AQUACULTURE TECHNOLOGY SECTOR: AN INDUSTRY ANALYSIS

• Main Author: Luh Putu Asri Cahyani, Ni
• Format: Theses
• Language: Indonesia
• Subjects: Manajemen umum
• Online Access: https://digilib.itb.ac.id/gdl/view/85761
• Institution: Institut Teknologi Bandung

This study investigates the application of machine learning techniques for predicting customer churn in the aquaculture technology sector. As the industry grows and competition intensifies, retaining customers becomes crucial for sustainable business growth. We utilize a synthesized dataset representative of the aquaculture technology industry to implement and evaluate several machine learning models, including Artificial Neural Networks, Decision Trees, Support Vector Machines, Random Forests, Logistic Regression, and gradient boosting techniques (XGBoost, LightGBM, and CatBoost). To address the common challenge of imbalanced datasets in churn prediction, we employ different data sampling strategies: SMOTE, SMOTE combined with Tomek Links, and SMOTE combined with Edited Nearest Neighbors. Additionally, we apply hyperparameter tuning to enhance model performance. Model evaluation is conducted using standard metrics such as Precision, Recall, F1-score, and the Receiver Operating Characteristic Area Under Curve (ROC AUC). The results demonstrate the potential of machine learning techniques in predicting customer churn in the aquaculture technology sector, with gradient boosting models showing superior performance. This study contributes to the understanding of customer behavior in the aquaculture technology sector and provides insights for developing targeted retention strategies in the industry.