1. An Adaptive Neural Network Model for Predicting Breast Cancer Disease in Mapped Nucleotide Sequences
- Author
-
Das, Lopamudra, Das, J. K., Nanda, Sarita, and Nanda, Sony
- Abstract
This article presents an adaptive neural network method for predicting breast cancer disease in mapped nucleotide sequences. It is based on the expansion of DNA input samples using an adaptive exponential functional link artificial neural network (AEFLANN). Exploiting the concept that the disease gene nucleotide sequence encompasses more randomness than the healthier ones, the discrimination is carried out by the magnitude of integral square error (ISE). The Least Mean Square/Fourth (LMS/F) algorithm is a potential candidate with faster convergence and reduced computational complexity that is used to predict the category of the genes taking into account the ISE as well as convergence characteristics. Comparing the performance of the proposed approach with the existing ones by using various evaluation parameters, it has been found that almost all the test genes have been segregated correctly within the first 50 iterations. Moreover, the proposed AE-FLANN-based LMS/F approach offers a classification accuracy of 96.5% for 65–35 training–testing partition and is also validated using tenfold cross-validation. Hence, the proposed approach is a suitable candidate that can be used to discriminate diseased genes from healthy ones.
- Published
- 2023
- Full Text
- View/download PDF