Breast carcinoma is a potential concern in medical field due to its violence and increased mortality rate. Early detection of all the significant abnormalities of the disease helps medical practitioners to start the treatment priorly for the increased survival rate of the patients. The advancements in engineering technology have shown that many soft computing techniques perform well for the early detection of the disease. Though the convolutional neural networks efficiently do the task, careful attention is needed in setting the hyperparameters for obtaining the optimal results. This article presents a very detailed view of the various machine learning algorithms used in diagnosis process of the breast carcinoma, implementation of those algorithms used, their performance comparison, CNN’s sequential model, its architecture, model’s performance with various optimizers and the result concluding the efficiency of the CNN model with various optimizers, hence providing a significant beneficial result for patients across the globe.