Scholarlink

Anaemia is a condition in which the red blood cell (RBC) oxygen-carrying ability is insufficient and is the most prevalent blood disorder globally. This disorder impacts quality of life as both an illness and a symptom. For the purpose of treating patients, accurate and timely identification of the kind of anaemia is essential. Such a diagnosis may be hampered by the growing patient population, hospital priorities, and challenges in contacting medical professionals. The current study suggests a method that will make it possible to identify anaemia in typical clinical practice settings. For this system, the model was created using 9 different machine-learning methods. Classification methods include bagging, Random Forest, Gradient boosting, Decision Tree, Extract trees, Support Vector Machines, kNeighbors and AdaBoost. A dataset of 1281 samples is utilized to evaluate the models, and 15 attributes are employed, including diagnosis, mean corpuscular volume (MCV), and mean corpuscular haemoglobin (MCH). Data are collected from https://www.kaggle.com . The interface’s purpose is to help students and medical consultants make decisions. Accuracy is attained for the nine distinct algorithms used to classify the data. The highest accuracy ( \(99.0\%\) ) was achieved using Bagging, followed by Random Forest ( \(98.9\%\) ), Gradient boosting ( \(98.8\%\) ), Decision Tree ( \(98.7\%\) ), Extract trees ( \(90.3\%\) ), Support Vector Machines ( \(79.1\%\) ), kNeighbors ( \(77.3\%\) ), Logistic regression ( \(0.76\%\) ), and AdaBoost ( \(48.9\%\) ). The one with the highest accuracy (Bagging) is validated Precision, Recall, F-score, support metrics, and Accuracy values.

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

Leveraging Machine Learning for Early and Accurate Anaemia Diagnosis: A Comparative Study of Classification Algorithms

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Oluwaseun Olumide Okundalaye,
  • Necati Özdemir,
  • Fırat Evirgen

摘要

Anaemia is a condition in which the red blood cell (RBC) oxygen-carrying ability is insufficient and is the most prevalent blood disorder globally. This disorder impacts quality of life as both an illness and a symptom. For the purpose of treating patients, accurate and timely identification of the kind of anaemia is essential. Such a diagnosis may be hampered by the growing patient population, hospital priorities, and challenges in contacting medical professionals. The current study suggests a method that will make it possible to identify anaemia in typical clinical practice settings. For this system, the model was created using 9 different machine-learning methods. Classification methods include bagging, Random Forest, Gradient boosting, Decision Tree, Extract trees, Support Vector Machines, kNeighbors and AdaBoost. A dataset of 1281 samples is utilized to evaluate the models, and 15 attributes are employed, including diagnosis, mean corpuscular volume (MCV), and mean corpuscular haemoglobin (MCH). Data are collected from https://www.kaggle.com . The interface’s purpose is to help students and medical consultants make decisions. Accuracy is attained for the nine distinct algorithms used to classify the data. The highest accuracy ( \(99.0\%\) ) was achieved using Bagging, followed by Random Forest ( \(98.9\%\) ), Gradient boosting ( \(98.8\%\) ), Decision Tree ( \(98.7\%\) ), Extract trees ( \(90.3\%\) ), Support Vector Machines ( \(79.1\%\) ), kNeighbors ( \(77.3\%\) ), Logistic regression ( \(0.76\%\) ), and AdaBoost ( \(48.9\%\) ). The one with the highest accuracy (Bagging) is validated Precision, Recall, F-score, support metrics, and Accuracy values.