Scholarlink

<p>Mycobacterium tuberculosis (MTB) detection is a major global health concern, requiring accurate and reliable diagnostic tools. According to the World Health Organization (WHO), the global burden of TB is staggering, with approximately 10 million new cases reported annually, leading to a devastating 1.5 million fatalities worldwide every year. The timely detection and treatment of TB are critical components in the fight against this disease, as they play a vital role in preventing its transmission, minimizing the risk of complications and deaths, and enhancing the overall quality of life for affected individuals. Hence, this study presents a novel Sea Lion Political Optimizer_Deep Maxout Network (SLnPO_DMN) for detecting MTB, leveraging a robust processing pipeline to achieve exceptional accuracy. The proposed pipeline involves a series of meticulous steps, including data pre-processing by Min–Max normalization to scale the data to a common range. To enhance model generalizability, data augmentation is carried out by Bootstrap, generating additional training samples by resampling the original data with replacement. The augmented data is then fed into a DMN, optimized by the SLnPO, a hybrid algorithm combining the strengths of sea lion optimization (SLnO) and political optimizer (PO). The SLnPO_DMN model is trained to identify and learn intricate patterns within the data, ultimately attaining optimal performance and demonstrating exceptional capability in detecting Mycobacterium tuberculosis with high accuracy and reliability. The SLnPO_DMN is evaluated using a comprehensive set of metrics, including accuracy, sensitivity, and specificity. The results demonstrate the effectiveness of the SLnPO_DMN model achieving an accuracy (99.12%), sensitivity of (98.74%), and specificity of (98%).</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

Refractive Index Biosensor–Based Detection of Mycobacterium Tuberculosis Using Sea Lion Political Optimizer and Deep Learning

  • 期刊论文

  • 发表日期：

    2025年4月30日

查看全文
  • V. K. Senthil Ragavan,
  • N. S. Nithya,
  • Anantharamaiah Vengala,
  • Balambigai Subramanian,
  • C. Ambhika

摘要

Mycobacterium tuberculosis (MTB) detection is a major global health concern, requiring accurate and reliable diagnostic tools. According to the World Health Organization (WHO), the global burden of TB is staggering, with approximately 10 million new cases reported annually, leading to a devastating 1.5 million fatalities worldwide every year. The timely detection and treatment of TB are critical components in the fight against this disease, as they play a vital role in preventing its transmission, minimizing the risk of complications and deaths, and enhancing the overall quality of life for affected individuals. Hence, this study presents a novel Sea Lion Political Optimizer_Deep Maxout Network (SLnPO_DMN) for detecting MTB, leveraging a robust processing pipeline to achieve exceptional accuracy. The proposed pipeline involves a series of meticulous steps, including data pre-processing by Min–Max normalization to scale the data to a common range. To enhance model generalizability, data augmentation is carried out by Bootstrap, generating additional training samples by resampling the original data with replacement. The augmented data is then fed into a DMN, optimized by the SLnPO, a hybrid algorithm combining the strengths of sea lion optimization (SLnO) and political optimizer (PO). The SLnPO_DMN model is trained to identify and learn intricate patterns within the data, ultimately attaining optimal performance and demonstrating exceptional capability in detecting Mycobacterium tuberculosis with high accuracy and reliability. The SLnPO_DMN is evaluated using a comprehensive set of metrics, including accuracy, sensitivity, and specificity. The results demonstrate the effectiveness of the SLnPO_DMN model achieving an accuracy (99.12%), sensitivity of (98.74%), and specificity of (98%).