An Early Cancer Prediction Based On Deep Neural Learning

Cancer is the one of the most leading cause of deaths among people The lifetime chance of acquiring this malignancy is one in eight for females. The prognosis for this condition greatly improves with early detection. Understanding the role of these genes in the disease is crucial. However, the complexity of gene data comes from there being too many aspects to study. Therefore, we need medical support from intelligent algorithms. Others have utilized prognostic approaches like machine-learning(ML) &deep-learning(DL) to foretell prognoses for tumor, and many researchers now use them to estimate the likelihood of patient survival. Predictive cancer prognosis accuracy is a hot topic right now. For accurate disease forecasting, deep neural learning (DNL) approaches are indispensable due to their speed and accuracy in making predictions based on massive amounts of clinical and genetic data. In this research, we present a DNLC for early cancer prediction based on deep neural learning. Experiments are run on the five cancer datasets to foretell the model's efficacy. It is capable of analyzing mammograms and picking out abnormalities that may be a sign of malignancy. To further enhance their predictive abilities, these frameworks can incorporate additional patient information, such as age and family records. Incorporating attention mechanisms into breast cancer prediction tasks is the focus of this research. Our proposed approach makes use of multimodal data to produce informative traits that enhance prognosis prediction for breast cancer. The first part of the model employs a convolutional neural network with sigmoid gated attention to generate stacked features, and the second part employs flatten, dense, and dropout operations to facilitate bi-modal attention. The experimental findings demonstrate that the proposed technique outperforms the conventional CNN and RNN models in terms of accuracy. Our results show that the DNLC method is the most effective overall, with an average accuracy of 95%.