Neural Network Model for Cancer Tumor Growth with Minimizing the Error in Learning

Abstract

Neural Network Learning is a process of exhibiting the original phenomena of any object in the network to simulate the desired output. Likewise, briefly processing the attributes of tumor identification from an ultrasound-screened image is a way to process it. A systematic approach for detecting malignant tumors is problematic approach. Given this fact, the data set used to classify the breast cancer tumor is less noisy. An ultrasound-screened image is a basic scanned image of a cancer tumor, which usually has the primary staging properties. Classification of malignant tumors in an ultrasound image is not as easy as possible for radiologists to predict the disease symptoms. Thus, automatic prediction is a challenging task for such images. Since it’s a distribution of artificial intelligence, this ability applies to cancer prediction. Automatic classification of benign and malignant in ultrasound images is bounded within the neural network. The features in a tumor image are learned with machine learning algorithms like K-Means and back propagation. Increasing the learning rate is essential for feature learning. The Sigmoid activation function is used for optimization and minimizing the error for feature learning. The ultrasound-screened image is of low quality in nature, and learning about features is a task in predicting.