Abstract:- With massiveimprovements in software technologies and artificial intelligence, nowadaysdata science is being used in multiple domains for various purposes. Deeplearning, machine learning, Artificial intelligence have gained prominence inrecent times and are successfully being used in domains like Speech Processing,Image Processing, Audio and Video processing, Text processing, natural languageprocessing, etc. The advantage of deep learning is that it has better learningabilities compared to other machine learning algorithms like supervisedlearning and semi-supervised learning. This feature of deep learning enables itto be used for different datasets for patterns and behavior extractions. Thecore pillar for developing a very good neural network is Perceptron. It is moreof a computational model and evaluates unsupervised data and thus it becomes atrusted tool for developing models for data analytics. The main objective ofthis project is to solve real-time problems like solar data classificationusing deep learning. In this project, we elaborate on how to construct deep neuralnetworks using perceptron and tensor flow. This methodology can be used tosolve many other real-time problems in multiple domains. In this project, wealso work with non-discrete datasets. We observe that our model is still wellbehaved and produced the desired results.
Keywords:- Deep Learning, TensorFlow , Machine Learning,,Perceptron Learning algorithm,