Scholarlink

<p>Diabetes is a major global health issue that requires efficient early detection and prediction technology. Predictive analytics and healthcare are explored in this paper, with a particular emphasis on diabetes prediction. It looks at the use of machine learning techniques, primarily using deep learning neural networks as the main focus. Developing reliable and effective models is a crucial problem for diabetes prediction. To address this, the study investigates a range of deep learning neural networks, including Feedforward Neural Networks (FNN), Recurrent Neural Networks (RNN), and Convolutional Neural Networks (CNN). These are the only models that can predict diabetes reliably based on many parameters such as body mass index (BMI), insulin levels, age, blood pressure, glucose levels, pregnancies, skin thickness, and diabetes pedigree function. Our extensive investigation’s findings indicated that the RNN model had the highest accuracy at 96.7%, while the FNN model placed in secondly at 95.6%. Both models demonstrated strong metrics for accuracy, recall, and F1-score, demonstrating their effectiveness in accurately forecasting the prevalence of diabetes. The accuracy rates of the CNN algorithm were 93.3% and 94.4%, respectively, showing impressive performance. The CNN model showed the fastest prediction time, at 0.107&#xa0;s, whereas the FNN model showed the fastest prediction time, at 0.087&#xa0;s, when we evaluated the prediction time efficiency of these models. Conversely, at 0.282&#xa0;s. These findings demonstrate how well deep learning models—in particular, RNN and FNN—can predict the occurrence of diabetes. Early diagnostic and management strategies in clinical and medical settings may be significantly impacted by this.</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

A Comparison of CNN, RNN, and FNN Algorithms to Investigate Effective Diabetes Prediction

  • 期刊论文

  • 发表日期：

    2025年12月2日

查看全文
  • B. P. Pradeep Kumar

摘要

Diabetes is a major global health issue that requires efficient early detection and prediction technology. Predictive analytics and healthcare are explored in this paper, with a particular emphasis on diabetes prediction. It looks at the use of machine learning techniques, primarily using deep learning neural networks as the main focus. Developing reliable and effective models is a crucial problem for diabetes prediction. To address this, the study investigates a range of deep learning neural networks, including Feedforward Neural Networks (FNN), Recurrent Neural Networks (RNN), and Convolutional Neural Networks (CNN). These are the only models that can predict diabetes reliably based on many parameters such as body mass index (BMI), insulin levels, age, blood pressure, glucose levels, pregnancies, skin thickness, and diabetes pedigree function. Our extensive investigation’s findings indicated that the RNN model had the highest accuracy at 96.7%, while the FNN model placed in secondly at 95.6%. Both models demonstrated strong metrics for accuracy, recall, and F1-score, demonstrating their effectiveness in accurately forecasting the prevalence of diabetes. The accuracy rates of the CNN algorithm were 93.3% and 94.4%, respectively, showing impressive performance. The CNN model showed the fastest prediction time, at 0.107 s, whereas the FNN model showed the fastest prediction time, at 0.087 s, when we evaluated the prediction time efficiency of these models. Conversely, at 0.282 s. These findings demonstrate how well deep learning models—in particular, RNN and FNN—can predict the occurrence of diabetes. Early diagnostic and management strategies in clinical and medical settings may be significantly impacted by this.