ABSTRACT
One of the main causes of death for women is breast cancer. Breast cancer is considered the most serious condition
affecting women. A study by IAFR revealed that globally, breast cancer has surpassed lung cancer as the most
prominent cancer found in women. Initial detection of this is extremely important, as it can save the patient's life
by preventing it from spreading further. Appropriate treatment can save lives because the fatality rate is very high
if it has spread to a significant area. Medical data is sophisticated, and even a small change in data can make a
huge difference. This needs accurate and powerful techniques that can detect even the smallest changes and
variations in the data as well as understand the underlying trends and patterns of the complex data. As the quantity
of data is increasing day by day, data-backed algorithms from ML and DL can be a game-changing factor in the
industry. This study brings both ML and DL techniques together to understand their performance on the Wisconsin
breast cancer dataset. We named our model Web-BCD. The main objective is to check the effectiveness and
efficiency of data classification, as measured by accuracy, precision, sensitivity, and specificity. After using more
than half a dozen ML techniques, the XGboost outperformed other techniques, while the DL method, which was
a simple neural network, was found to be prone to overfitting breast cancer data. It performed terribly and gave
only 37% accuracy on test data.