Discovery of Hidden Pattern in Thyroid Disease by Machine Learning Algorithms

Background: Decision tree provides help in making decision for very complex and large dataset. Decision tree techniques are used for gathering knowledge. Classification tree algorithms predict the experimental values of women thyroid dataset. The objective of this research paper observation is to determine hyperthyroidism, hypothyroidism and euthyroidism participation in hormones can be good predictor of the final result of laboratories and to examination whether the propose ensemble approach can be similar accuracy to other single classification algorithm.

Results: In the proposed experiment real data from 499 thyroid patients were used classifications algorithms in predicting whether thyroid detected or not detected on the basis of T3, T4 and TSH experimental values. The results show that the expectation of maximization classification tree algorithms in those of the best classification algorithm especially when using only a group of selected attributes. Finally we predict batch size, tree confidential factor, min number of observation, num folds, seed, accuracy and time build model with different classes of thyroid sickness.

Conclusion: Different classification algorithms are analyzed using thyroid dataset. The results obtained by individual classification algorithms like J48, Random Tree and Hoeffding gives accuracy 99.12%, 97.59% and 92.37 respectively. Then we developed a new ensemble method and apply again on the same dataset, which gives a better accuracy of 99.2% and sensitivity of 99.36%. This new proposed ensemble method can be used for better classification of thyroid patients.