A multi-layer perceptron based improved thyroid disease prediction system

Arvind Selwal, Ifrah Raoof


A challenging task for the medical science is to achieve the accurate diagnosis of diseases prior to its treatment. A pattern classifier is used for solving complex and non-separable computing problems in different fields like biochemical analysis, image processing and chemical analysis etc .The accuracy for thyroid diagnosis system may be improved by considering few additional attributes like heredity ,age, anti-bodies etc.  In this paper, a thyroid disease prediction system is developed using multilayer perceptron (MLP). The proposed system uses 7–11 attributes of individuals to classify them in normal, hyperthyroid and hypothyroid classes. The proposed model uses gradient descent backpropogation algorithm for training the multilayer perceptron using dataset of 120 subjects. The thyroid prediction system promises excellent overall accuracy of ~100% for 11 attributes. However, the system results in a lower accuracy of 66.7% using 11 attributes and 70% using 7 attributes with 30 subjects.


Multilayer perceptron; pattern classifier; thyroid disease; back-propogation; learning

Full Text:


DOI: http://doi.org/10.11591/ijeecs.v17.i1.pp524-532


  • There are currently no refbacks.

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

shopify stats IJEECS visitor statistics