Scholarlink

<p>Most of the countries around the world depends on agriculture for their growth and development. There are different types of agricultural methods, irrigated agriculture is one of them. In irrigated agriculture fresh water acts as the main component. With the progress of time, the scarcity of fresh water is noticed around the globe and it will become worse in upcoming days. Now-a-days to get rid off such an alarming issue, the only workable alternatives are precision agriculture and intelligence-based irrigation. Intelligent-based irrigation and precision agriculture have only recently become financially feasible with rise of the Machine Learning (ML) and Internet-of-Things (IoT). IoT has several advantages such as improved productivity, reduced costs, gain in energy, prediction of events, and enhancement of comfort in different fields of human society. Being emerged with different technologies and information handling aids, IoT becomes susceptible to be hampered from security and confidentiality perspective. In this paper, a methodology is discussed for identifying and categorizing different attacks into IoT-based agricultural paradigm. In all IoT areas, including those linked to agriculture, security and confidentiality are top priorities. The NSL-KDD dataset is utilized in this work, which is initially preprocessed by converting symbolized attributes into numerical ones. Linear Discriminant Analysis (LDA), Principal Component Analysis (PCA) and T-Distributed Stochastic Neighbour Embedding (T-SNE) are applied to extract features. Subsequently, the preprocessed dataset undergoes classification utilizing ML algorithms like Logistic Regression (LR), Support Vector Machine (SVM), K-Nearest Neighbour (KNN), Random Forest (RF), Gaussian Naive Bayes (GNB), Decision Tree (DT), and Stochastic Gradient Descent (SGD). The suggested approach outperforms existing methods, as evidenced by precision, accuracy, recall, F1-Score, Area Under the ROC Curve (AUC), False Discovery Rate (FDR) and Critical Success Index (CSI).</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

RETRACTED ARTICLE: Reinforcing internet of things security measures in smart agriculture with feature extraction-based supervised machine learning for intrusion detection

  • 期刊论文

  • 发表日期：

    2025年4月24日

查看全文
  • Sarbajit Manna,
  • Arindam Sarkar,
  • Atanu Mondal

摘要

Most of the countries around the world depends on agriculture for their growth and development. There are different types of agricultural methods, irrigated agriculture is one of them. In irrigated agriculture fresh water acts as the main component. With the progress of time, the scarcity of fresh water is noticed around the globe and it will become worse in upcoming days. Now-a-days to get rid off such an alarming issue, the only workable alternatives are precision agriculture and intelligence-based irrigation. Intelligent-based irrigation and precision agriculture have only recently become financially feasible with rise of the Machine Learning (ML) and Internet-of-Things (IoT). IoT has several advantages such as improved productivity, reduced costs, gain in energy, prediction of events, and enhancement of comfort in different fields of human society. Being emerged with different technologies and information handling aids, IoT becomes susceptible to be hampered from security and confidentiality perspective. In this paper, a methodology is discussed for identifying and categorizing different attacks into IoT-based agricultural paradigm. In all IoT areas, including those linked to agriculture, security and confidentiality are top priorities. The NSL-KDD dataset is utilized in this work, which is initially preprocessed by converting symbolized attributes into numerical ones. Linear Discriminant Analysis (LDA), Principal Component Analysis (PCA) and T-Distributed Stochastic Neighbour Embedding (T-SNE) are applied to extract features. Subsequently, the preprocessed dataset undergoes classification utilizing ML algorithms like Logistic Regression (LR), Support Vector Machine (SVM), K-Nearest Neighbour (KNN), Random Forest (RF), Gaussian Naive Bayes (GNB), Decision Tree (DT), and Stochastic Gradient Descent (SGD). The suggested approach outperforms existing methods, as evidenced by precision, accuracy, recall, F1-Score, Area Under the ROC Curve (AUC), False Discovery Rate (FDR) and Critical Success Index (CSI).