Scholarlink

The endocrine system’s thyroid gland generates hormones that control the body’s metabolism, growth, development, excess or under secretion of thyroid causes complications. There are mainly 2 types of thyroid disease—hyperthyroidism and hypothyroidism. In hypothyroidism thyroid gland doesn’t make enough thyroid hormone. Another name for this illness is underactive thyroid. Hyperthyroidism is the cause of thyroid gland producing large amount of thyroid hormones. This condition also is called overactive thyroid. To detect thyroid disease, blood tests and medical imaging are performed (ultrasound). Machine learning algorithms are most useful in predicting hypothyroidism in patients. Various types of classification algorithms like support vector machine, decision tree, logistic regression were used for the data set. Accuracy, precision, F1 score of algorithms were considered. We have obtained accuracy of 99.6% in decision tree, 94.8% in random forest classifier, 88.7% in SVM, 95.75% in KNN classifier, 96.5% in logistic regression. We have obtained precision of 100% in decision tree, 24.6% in random forest classifier, 18.46% in SVM, 49.2% in KNN classifier, 53.8% in logistic regression. We have obtained F1 score of 97.7% in decision tree, 39.5% in random forest classifier, 18.4% in SVM, 61.5% in KNN classifier, 67.9% in logistic regression. And also we have calculated the percentage of recall, specificity, area under curve-AUC, receiver operating characteristic-ROC. We have obtained recall score of 95.58% in decision tree, 100% in random forest classifier, 82.05% in KNN, 18.46% in SVM, 92.01% in logistic regression. We have obtained of 95.58% in decision tree, 100% in random forest classifier, 82.05% in KNN, 18.46% in SVM, 92.01% in logistic regression. Recall score of each algorithm obtained are decision tree classifier—95.5%, random forest classifier—5.1%, KNN classifier—82%, SVM—18.4%, logistic regression—92.1%. Specificity score obtained are decision tree classifier—100%, random forest classifier—94.6%, KNN classifier—96.3%, SVM—93.9%, logistic regression—96.6%. We have obtained a roc curve as decision tree classifier—95.5%, random forest classifier—5.1%, KNN classifier—82%, SVM—18.4%, logistic regression—92.1%. Displaying of the result has been showcased using Graphical User Interface.

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

Blockchain-Based Healthcare Recommender System Using Deep Learning

  • 章节

  • 发表日期：

    2025年

查看全文
  • J. L. Hamsapriya,
  • Tharanga Shetty,
  • B. Neha,
  • G. Pallavi,
  • C. Gururaj

摘要

The endocrine system’s thyroid gland generates hormones that control the body’s metabolism, growth, development, excess or under secretion of thyroid causes complications. There are mainly 2 types of thyroid disease—hyperthyroidism and hypothyroidism. In hypothyroidism thyroid gland doesn’t make enough thyroid hormone. Another name for this illness is underactive thyroid. Hyperthyroidism is the cause of thyroid gland producing large amount of thyroid hormones. This condition also is called overactive thyroid. To detect thyroid disease, blood tests and medical imaging are performed (ultrasound). Machine learning algorithms are most useful in predicting hypothyroidism in patients. Various types of classification algorithms like support vector machine, decision tree, logistic regression were used for the data set. Accuracy, precision, F1 score of algorithms were considered. We have obtained accuracy of 99.6% in decision tree, 94.8% in random forest classifier, 88.7% in SVM, 95.75% in KNN classifier, 96.5% in logistic regression. We have obtained precision of 100% in decision tree, 24.6% in random forest classifier, 18.46% in SVM, 49.2% in KNN classifier, 53.8% in logistic regression. We have obtained F1 score of 97.7% in decision tree, 39.5% in random forest classifier, 18.4% in SVM, 61.5% in KNN classifier, 67.9% in logistic regression. And also we have calculated the percentage of recall, specificity, area under curve-AUC, receiver operating characteristic-ROC. We have obtained recall score of 95.58% in decision tree, 100% in random forest classifier, 82.05% in KNN, 18.46% in SVM, 92.01% in logistic regression. We have obtained of 95.58% in decision tree, 100% in random forest classifier, 82.05% in KNN, 18.46% in SVM, 92.01% in logistic regression. Recall score of each algorithm obtained are decision tree classifier—95.5%, random forest classifier—5.1%, KNN classifier—82%, SVM—18.4%, logistic regression—92.1%. Specificity score obtained are decision tree classifier—100%, random forest classifier—94.6%, KNN classifier—96.3%, SVM—93.9%, logistic regression—96.6%. We have obtained a roc curve as decision tree classifier—95.5%, random forest classifier—5.1%, KNN classifier—82%, SVM—18.4%, logistic regression—92.1%. Displaying of the result has been showcased using Graphical User Interface.