Breast Cancer Diagnosis with Machine Learning Using Feed-Forward Multilayer Perceptron Analog Artificial Neural Network

Abstract

An analog network classifier based on a multiplier and non-linear functions is presented in this paper, executing binary classification on breast cancer cells, and categorizing biopsies as benign or malignant tumors. An off-chip learning on-chip inference methodology is proposed for implementing a feed-forward analog artificial neural network based on fundamental design analog block circuits, realized with the aid of 90 nm CMOS technology. These circuits are meticulously designed and fine-tuned at the transistor scale to meet design criteria while minimizing power consumption. Through Spice simulations, the basic analog blocks were developed, leading to the specification of the full-chip hardware neural network. The Monte Carlo analysis of the final circuit reveals that the network achieves 96.85% accuracy and 0.9309 MCC on the Wisconsin breast cancer dataset, with a power consumption of 31.95 μW, and power supply rail of ±900 mV per analog circuit component and computational unit. The model effectively captures data patterns, providing stable, reliable, and robust predictions.