Artificial Neural Network-Based Leukaemia Identification and Prediction using Ensemble Deep Learning Model

Abstract

Leukemia (ALL) is a type of blood cancer that causes a huge number of deaths throughout the Globe. Technology is advancing in decision making like moving from manual inspection to automatic (using deep learning) detection. As flaws persist in manual identification, detection through deep ensemble learning on enhanced augmented images or datasets led to flawless identification. Here two different datasets were used from Kaggle. The proposed artificial neural network on classified data by ensemble classifier led to the generation of best accuracy. The proposed method can provide 100% accuracy with a good quality dataset. Whereas with poor quality data set also proposed method can provide 96.3% of accuracy. Elapsed time for the best-case dataset is 0.366137 whereas 0.38861 for the worst-case dataset. The mean square error is 0.00911. It is analyzed with both types of datasets, producing efficient results.