Enhancing Prediction of Sonar Rock with Mines Comparison of Diverse Machine Learning Algorithms

  • Unique Paper ID: 166328
  • Volume: 11
  • Issue: 2
  • PageNo: 594-599
  • Abstract:
  • Underwater navigation poses significant challenges, one of which is accurately differentiating rocks from mines using passive sonar. This study investigates the efficacy of four machine learning algorithms in this task: K-Nearest Neighbors (KNN), Support Vector Machines (SVM), Decision Trees, and Random Forest. Employing a benchmark sonar dataset containing extracted features, we evaluated their performance in classifying rocks and mines. Random Forest emerged as the most accurate model, demonstrating the strengths of ensemble learning in reducing overfitting and boosting accuracy. While SVM delivered comparable performance, Decision Trees and KNN exhibited slightly lower efficacy. These findings underscore the potential of machine learning, particularly Random Forest, for enhancing sonar-based rock and Mine classification. Future research avenues include exploring advanced feature engineering and hyperparameter optimization, along with delving into interpretable models and active learning methodologies. This research paves the way for improved underwater navigation, safety, and operational efficiency, contributing significantly to safer and more effective ocean exploration and utilization. This research highlights the promising potential of machine learning for real-world applications like underwater object detection, paving the way for safer and more efficient ocean exploration and operations.

Cite This Article

  • ISSN: 2349-6002
  • Volume: 11
  • Issue: 2
  • PageNo: 594-599

Enhancing Prediction of Sonar Rock with Mines Comparison of Diverse Machine Learning Algorithms

Related Articles