Performance Analysis of Classification Algorithms for Dry Bean Prediction
Bist, Mohan Bahadur (2024)
2024
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2024060320079
https://urn.fi/URN:NBN:fi:amk-2024060320079
Tiivistelmä
This thesis studies the efficacy of machine learning in dry bean classification. The Random Forest (RF), Optimized Forest (OF), and Logistic Model Tree (LMT) algorithms were evaluated using an openly accessible dataset and k-fold cross-validation. The study aimed to determine the most accurate and reliable approach for predicting dry bean types.
All three algorithms obtained good accuracy, with performance varied according to the number of folds used in cross-validation (k = 5, 10, and 15). In the 5-fold arrangement, the Logistic Model Tree had the highest accuracy (92.46%). The Random Forest performed best in the 10-fold setting (92.59% accuracy), whereas the Optimized Forest proved to be the most reliable option across all folds (92.65% accuracy in the 15-fold arrangement).
These findings indicate that machine learning has the potential to automate dry bean sorting, improve quality control, and promote precision agriculture. Future research might focus on enhancing computational efficiency, studying deep learning algorithms, and creating real-time field decision-making systems.
All three algorithms obtained good accuracy, with performance varied according to the number of folds used in cross-validation (k = 5, 10, and 15). In the 5-fold arrangement, the Logistic Model Tree had the highest accuracy (92.46%). The Random Forest performed best in the 10-fold setting (92.59% accuracy), whereas the Optimized Forest proved to be the most reliable option across all folds (92.65% accuracy in the 15-fold arrangement).
These findings indicate that machine learning has the potential to automate dry bean sorting, improve quality control, and promote precision agriculture. Future research might focus on enhancing computational efficiency, studying deep learning algorithms, and creating real-time field decision-making systems.