Akademik Veri Yönetim Sistemi
6th International Applied Statistics Congress, Ankara, Türkiye, 14 - 16 Mayıs 2025, ss.647-653, (Tam Metin Bildiri)
This research focuses on predicting plant growth milestones using a dataset containing keyenvironmental and agricultural parameters such as soil type, sunlight duration, irrigation frequency,fertilizer type, temperature, and humidity. Several supervised machine learning models were appliedto classify plant growth stages: Gaussian Naive Bayes, Bernoulli Naive Bayes, Logistic Regression, k-Nearest Neighbors, Decision Tree, Random Forest, Support Vector Classifier, and Deep NeuralNetwork. A threshold optimization technique was used to fine-tune the classification results. Amongall the models, Random Forest Classifier provided the best overall performance with an accuracy of91.0% and a ROC AUC score of 0.9624. Despite being more computationally intensive, Deep NeuralNetwork also produced promising results in terms of precision and F1 score. ROC AUC curves wereconstructed to demonstrate the comparative effectiveness of the models. The findings highlight thepotential of both classical and deep learning approaches in modeling plant growth, providingmeaningful support for agricultural planning and greenhouse management.