International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 09 Issue: 08 | Aug 2022
p-ISSN: 2395-0072
www.irjet.net
Comparative Analysis of Early Detection of Hypothyroidism using Machine Learning Techniques Ranjitha B1, K R Sumana2 1PG
Student, The National Institute of Engineering, Mysuru, Karnataka, India 2Assistant Professor, Mysuru, Karnataka, India ---------------------------------------------------------------------***--------------------------------------------------------------------hypothyroidism, and that this threat increased with longer Abstract - The diagnosis of health conditions and proper
employment as a night shift worker. While sleeping further than eight hours per day may increase the threat of both hyperactive and underactive thyroid function. Thyroid disease affect an estimated 200 million people worldwide. In India there are 42 Million people have thyroid diseases and Hypothyroidism is utmost of the common thyroid complaint in India.
treatment of disease at an early stage is one of the most challenging tasks in the healthcare field. Hypothyroidism is a type of thyroid disease. Thyroid glands are located in the middle of our necks. It has a butterfly shape and is small in size. People with hypothyroidism do not produce enough thyroid hormone to keep their bodies functioning normally. The thyroid gland may be involved in several conditions either directly or indirectly. Damage to the thyroid gland and inflammation are the causes of hypothyroidism. Low thyroid hormone levels cause the body’s functions to slow down, leading to general symptoms like fatness, low pulse, increase in cold sensitiveness, neck swelling, dry skin, hands symptom, hair drawback, serious emission periods. The purpose of this project is to predict the Hypothyroidism disease at the early stage. Nowadays, machine learning has become an incredibly popular way to detect various diseases. Machine learning is used to detect disease at an early stage with greater accuracy. This Project uses KNN, Random Forest(RF) and XGB algorithms to predict the hypothyroidism disease at the early stage.
2. RELATED WORK The authors [1] in this article applied the classification (KNN) and prediction model (decision tree) to the thyroid dataset to accurately predict new patient entry. The KNN algorithm is used to classify thyroid disorders with related prioritized symptoms. Artificial Neural Network, support vector machine, Naive Bayes and KNearest Neighbor are the important modes applied to the prediction of thyroid disease and the results show that the K-nearest neighbor accuracy is better than any other thyroid disease detection technique. [2]utilized information mining calculations, for example, KNN, Naive Bayes, Support Vector Machine for the concentrate in this paper. The after effects of these arrangement techniques depend on the precision and execution of the model. For the given dataset, SVM accuracy is 0.82, Naive Bayes accuracy is 0.83 and KNN accuracy is 0.85. [3] Utilizes calculations like KNN, Random Forest, Naive Bayes, and ANN. KNN with Random Forest exhibited improved results with a precision of 94.8 percent when contrasted with the complete outcomes with four classifiers on the equivalent dataset. Utilizing decision tree algorithm, random forest algorithm, support vector machine algorithm, logistic regression and multilayer feedforward algorithm[4]. After doing a comparative analysis to identify the prediction algorithm that produces the most precise and accurate results, it can be said that the decision tree algorithm does so with a 99.46 percent accuracy rate and precision 0.99. The informational collections for the thyroid sicknesses have been had from the UCI website. The Machine Learning Algorithms like Artificial Neural Network, Support Vector Machine, Decision Tree, K-Nearest Neighbor are utilized to arrange and anticipate the exactness. Thyroid infection prescient models which require least number of boundaries of an individual to analyze thyroid illness and sets aside both cash and season of the patient. [5]This paper studies on thyroid disease and apply some algorithms to test performance study on mentioned algorithms. ANN-97.50,
Key Words: Thyroid disease, Hypothyroidism, KNN, Random Forest ,XGB
1.INTRODUCTION Thyroid is one of our glands, which make hormones. Thyroid hormones control the rate of numerous conditioning in our body. It secretes a few chemicals that are blended in with blood and excursion across the body to control modling. There are two primary thyroid chemicals Triiodothyronine( T3) and Thyroxin( T4). These two chemicals are significantly answerable for keeping up with the energy in our bodies. The two main types of thyroid condition are Hypothyroidism and Hyperthyroidism. Hypothyroidism is caused when the gland releases low situations of thyroid hormone. Symptoms of an underactive thyroid(hypothyroidism) can include Feeling tired, Gaining weight, passing obliviousness, Having frequent and heavy menstrual ages, Having dry and coarse hair, Having a coarse voice. Foods that affect thyroid are tofu, tempeh, edamame sap, soy milk, etc. The potables coffee, green tea, and alcohol — these potables may irritate our thyroid gland. People who worked 53– 83 hours per week were shown to have a higher rate of hypothyroidism than those who worked 36–42 hours per week. The night shift work might be associated with the threat of subclinical
© 2022, IRJET
|
Impact Factor value: 7.529
|
ISO 9001:2008 Certified Journal
|
Page 230