Neural Networks – Classification of Tumors

In this collaborative research project, my team and I rigorously examined the capabilities of neural networks in classifying tumours as malignant or benign. Utilizing the "Wisconsin Breast Cancer Original" dataset, we designed a neural network model with nine input layers representing various tumour attributes and two output layers for the classification results.

Our study employed Matlab's Neural Net Toolkit to fine-tune the model by altering parameters like the number of hidden neurons and the ratios of training, validation, and testing data. Our findings reveal that the optimal balance between these parameters is crucial for achieving accurate and efficient classification. The research is a comprehensive guide for leveraging neural networks in medical diagnostics, particularly oncology.