Scholarlink

<p>An advancing neurological condition that impairs motor control is Parkinson’s disease. It is defined by the slow deterioration of some brain neurons, or nerve cells, especially those responsible for the production of dopamine, a neurotransmitter that is essential for coordinating motor functions. Parkinson’s disease progresses dynamically and modifies over time. Accurate predictions may require ongoing monitoring since machine learning models may find it difficult to adjust to these temporal fluctuations. The proposed paper uses the Explainable AI (XAI) framework, to offer comprehensible and clear justifications for their choices and deeds. Healthcare experts can evaluate the performance of the model and provide insightful feedback for its ongoing improvement using the explanations offered by XAI. To ensure AI models are accurate in real-world clinical contexts, iterative feedback loops are crucial for their improvement. The paper implements LIME (Local Interpretable Model-agnostic Explanations) for classification of disease after performing feature selection using mutual information and pre-processing using synthetic data generator. To assist users in understanding which input features have the greatest influence on the model’s decision-making, LIME offers a rating of feature importance for a particular instance. The proposed model has achieved 98.75%.</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

Explainable AI in parkinson’s disease detection: leveraging lime for transparent model interpretation and improvement

  • 期刊论文

  • 发表日期：

    2025年9月8日

查看全文
  • K. Anji Reddy,
  • Syed Ziaur Rahman,
  • B. Manikyala Rao,
  • Mohammed Ali Hussain,
  • C. V. K. Sirisha,
  • Malathi Puligadda

摘要

An advancing neurological condition that impairs motor control is Parkinson’s disease. It is defined by the slow deterioration of some brain neurons, or nerve cells, especially those responsible for the production of dopamine, a neurotransmitter that is essential for coordinating motor functions. Parkinson’s disease progresses dynamically and modifies over time. Accurate predictions may require ongoing monitoring since machine learning models may find it difficult to adjust to these temporal fluctuations. The proposed paper uses the Explainable AI (XAI) framework, to offer comprehensible and clear justifications for their choices and deeds. Healthcare experts can evaluate the performance of the model and provide insightful feedback for its ongoing improvement using the explanations offered by XAI. To ensure AI models are accurate in real-world clinical contexts, iterative feedback loops are crucial for their improvement. The paper implements LIME (Local Interpretable Model-agnostic Explanations) for classification of disease after performing feature selection using mutual information and pre-processing using synthetic data generator. To assist users in understanding which input features have the greatest influence on the model’s decision-making, LIME offers a rating of feature importance for a particular instance. The proposed model has achieved 98.75%.