Scholarlink

Lumpy Skin Disease (LSD) is a highly contagious viral disease affecting cattle, necessitating early and accurate detection for effective disease control. This study explores the application of automated machine learning (AutoML) and deep learning (DL) techniques for LSD classification using epidemiological and geographical data. We evaluate multiple models, including TPOT (AutoML), Deep Neural Networks (DNNs), Long Short-Term Memory (LSTM), Bidirectional LSTM (BILSTM), Gated Recurrent Units (GRUs), Bidirectional GRU (BIGRU), and Deep Belief Networks (DBNs), across two train–test split ratios (80–20 and 70–30), validated with KFold and StratifiedKFold cross validation. To enhance model transparency, Local Interpretable Model-Agnostic Explanations (LIME) is applied to the best-performing model, providing insights into feature importance and improving trust in AI-driven diagnostics.

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

Enhancing the Explainability of AutoML and Deep Learning Techniques for Lumpy Skin Disease Detection Using Model Agnostic Explainable AI

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Aadrian Routh,
  • Tapan Kumar Dey,
  • Eesh Khanna,
  • Onkar Singh

摘要

Lumpy Skin Disease (LSD) is a highly contagious viral disease affecting cattle, necessitating early and accurate detection for effective disease control. This study explores the application of automated machine learning (AutoML) and deep learning (DL) techniques for LSD classification using epidemiological and geographical data. We evaluate multiple models, including TPOT (AutoML), Deep Neural Networks (DNNs), Long Short-Term Memory (LSTM), Bidirectional LSTM (BILSTM), Gated Recurrent Units (GRUs), Bidirectional GRU (BIGRU), and Deep Belief Networks (DBNs), across two train–test split ratios (80–20 and 70–30), validated with KFold and StratifiedKFold cross validation. To enhance model transparency, Local Interpretable Model-Agnostic Explanations (LIME) is applied to the best-performing model, providing insights into feature importance and improving trust in AI-driven diagnostics.