Multi-class Weather Classification: Comparative Analysis of Machine Learning Algorithms

Weather is one of the events that have a huge impact on our society. Often, the weather tends to get a bit difficult and troubling for us making predictions a bit challenging. The current system uses an array of sensors to get the correct weather status. In this paper, we have implemented weather classification using image data. The use of many classifiers has been shown, and it also has a comparative analysis between the classifiers which will give out a better understanding of the techniques used for creating classifiers. The techniques used for implementing these classification models are machine learning and deep learning. Used models are as follows, support vector machine (SVM), K-nearest neighbor (KNN), neural networks (Perceptron), and convolutional neural network (CNN). The research paper will show you the time taken to build a model, its accuracy along with few other performance metrics, to test out the efficiency of our models. The convolutional neural network model, which is based on deep learning concepts and is well suited for visual analysis, is the best of them all.