Scholarlink

<p>Rice growth is highly reliant on irrigation water due to the uneven distribution of rainfall. Scheduling irrigation depends on soil water balance, or soil moisture (SM), because if SM is lower, irrigation water requirements will be higher, and vice versa. Adequate soil water balance ensures optimal rice growth and also maximizes production. This study aims to develop an accurate SM prediction model using Random Forest optimized by Particle Swarm Optimization (PSO-RF), Decision Tree optimized by PSO (PSO-DT), and K-Nearest Neighbor optimized by PSO (PSO-K-NN). A modified Binary Ant Colony Optimization (mBACO) algorithm is employed to select correlated weather and soil variables in conjunction with SM. The Binary Ant Colony Optimization algorithm is modified by applying the Cauchy distribution for speedy convergence and Levy mutation to escape multiple local optima and for better accuracy in feature selection performance. Accurate feature selection reduces the overfitting problem, boosts the prediction accuracy, and reduces the training time of the prediction algorithms. Experimental results show that PSO-RF outperforms other standalone ML algorithms and other PSO-optimized prediction algorithms in terms of performance evaluation metrics (accuracy <InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="42979_2025_4162_Article_IEq1.gif" Format="GIF" Height="16" Rendition="HTML" Resolution="72" Type="Linedraw" Width="31" /> </InlineMediaObject> <EquationSource Format="TEX">\(95\%\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mn>95</mn> <mo>%</mo> </mrow> </math></EquationSource> </InlineEquation>, MAE 0.14, MSE 0.12, RMSE 0.34, Precision, Recall, and F1-score of 94.52, 94.95 and 94.74) respectively with correlated features selected using mBACO algorithm. An automated cost-effective irrigation approach is then proposed using a Fuzzy logic controller for improving the rice yield based on accurate SM prediction.</p>

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

Optimized Soil Moisture Prediction Using RF-PSO and Modified BACO Based Feature Selection

  • 期刊论文

  • 发表日期：

    2025年7月18日

查看全文
  • Parijata Majumdar,
  • Sanjoy Mitra

摘要

Rice growth is highly reliant on irrigation water due to the uneven distribution of rainfall. Scheduling irrigation depends on soil water balance, or soil moisture (SM), because if SM is lower, irrigation water requirements will be higher, and vice versa. Adequate soil water balance ensures optimal rice growth and also maximizes production. This study aims to develop an accurate SM prediction model using Random Forest optimized by Particle Swarm Optimization (PSO-RF), Decision Tree optimized by PSO (PSO-DT), and K-Nearest Neighbor optimized by PSO (PSO-K-NN). A modified Binary Ant Colony Optimization (mBACO) algorithm is employed to select correlated weather and soil variables in conjunction with SM. The Binary Ant Colony Optimization algorithm is modified by applying the Cauchy distribution for speedy convergence and Levy mutation to escape multiple local optima and for better accuracy in feature selection performance. Accurate feature selection reduces the overfitting problem, boosts the prediction accuracy, and reduces the training time of the prediction algorithms. Experimental results show that PSO-RF outperforms other standalone ML algorithms and other PSO-optimized prediction algorithms in terms of performance evaluation metrics (accuracy \(95\%\) 95 % , MAE 0.14, MSE 0.12, RMSE 0.34, Precision, Recall, and F1-score of 94.52, 94.95 and 94.74) respectively with correlated features selected using mBACO algorithm. An automated cost-effective irrigation approach is then proposed using a Fuzzy logic controller for improving the rice yield based on accurate SM prediction.