| Autor: |
Kumar, Jakkulla Pradeep, Muppagowni, Ganesh Karthik, Kumar, Jayapal Praveen, Malla, Sree Jagadeesh, Chandanapalli, Suresh Babu, Sandhya, Ethala |
| Předmět: |
|
| Zdroj: |
Ingénierie des Systèmes d'Information; Apr2023, Vol. 28 Issue 2, p299-308, 10p |
| Abstrakt: |
Thyroid is on the rise all across the world in modern times. The prevalence of thyroid disease in India is notably high, reaching 1 in 10. Due to the general public's lack of knowledge, the situation with that illness is fast deteriorating. Early diagnosis is crucial so that medical professionals can administer effective treatment before the condition worsens. This is especially true when using deep learning (DL) to predict sickness. One of DL's strengths is its ability to predict how a disease will progress in the future. Once more, several feature selection procedures have benefited in the process of disease prediction and assumption. The most common types of hypothyroidism in this study, we make an effort to predict the initial stage of thyroid development. To achieve this goal, the research has relied heavily on the feature selection strategy in addition to several different categorization methods. Each iteration of the projected adaptive tunicate swarm optimisation (ATSA) consists of two primary phases: searching all over the search space using an arbitrarily picked tunicate and refining the search using the position of the finest tunicate. By making this adjustment, the procedure is better able to explore its environment while simultaneously being protected from the dangers of a sudden convergence. Additionally, a deep convolutional neural network (DeepCNN) is used for disease identification, and the Grey Wolf Optimizer (GWO) is used for its training. Both could be associated more accurately. We were able to improve the suggested model's accuracy to 95% after tweaking the dataset, with 92% specificity. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
