Scholarlink

The present study investigates the application of Artificial Neural Networks (ANN) to address the problem of dissipative coefficient (CD) prediction in the modeling of the wave-vegetation interaction process. The key parameters involved in the wave-vegetation studies were identified through systematic correlation studies from the experimental datasets. Pearson, Spearman, and Principal Component Analysis (PCA) were used to determine the contribution of the input features toward CD prediction for the ANN model. Based on the understanding from our experimental studies, the modified submergence ratio, single-cylinder inertial coefficient ( \(C_{m}^{*}\) ), and drag coefficient ( \(C_{d}^{*}\) ) are included in the input features from the literature. The correlation coefficient and PCA analysis suggest that these features contribute significantly to the overall variance in the experimental datasets. This provides motivation to further investigate their potential impact on the predictive model and assess whether incorporating them enhances model performance. The machine learning (ML) approach was leveraged to analyze the effect of including these additional features as the model inputs in the prediction of dissipative coefficient CD. This study shows the performance and effect of these features included in the model prediction compared to the existing empirical model. Furthermore, this model can be a potential tool supporting the macroscopic numerical studies related to wave-vegetation interaction studies.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Machine Learning Approach Toward Engineering Vegetation as Coastal Protection Measures

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • N. Hari Ram,
  • S. Sangeetha,
  • V. Sriram,
  • Abhilash Somayajula

摘要

The present study investigates the application of Artificial Neural Networks (ANN) to address the problem of dissipative coefficient (CD) prediction in the modeling of the wave-vegetation interaction process. The key parameters involved in the wave-vegetation studies were identified through systematic correlation studies from the experimental datasets. Pearson, Spearman, and Principal Component Analysis (PCA) were used to determine the contribution of the input features toward CD prediction for the ANN model. Based on the understanding from our experimental studies, the modified submergence ratio, single-cylinder inertial coefficient ( \(C_{m}^{*}\) ), and drag coefficient ( \(C_{d}^{*}\) ) are included in the input features from the literature. The correlation coefficient and PCA analysis suggest that these features contribute significantly to the overall variance in the experimental datasets. This provides motivation to further investigate their potential impact on the predictive model and assess whether incorporating them enhances model performance. The machine learning (ML) approach was leveraged to analyze the effect of including these additional features as the model inputs in the prediction of dissipative coefficient CD. This study shows the performance and effect of these features included in the model prediction compared to the existing empirical model. Furthermore, this model can be a potential tool supporting the macroscopic numerical studies related to wave-vegetation interaction studies.