Most research on Carnatic classical music has focused on identifying ragas based on melody and rhythm. However, no study has been carried out to classify ragas into different types: Vivadhi ragas and Avivadhi ragas. A Raga is said to be Vivadhi if the two notes rendered are closer in frequency, resulting in a discordant effect, and if such an effect is not produced, it is said to be an Avivadhi Raga. Ragas can be classified based on the presence of certain notes and the effect generated by how one note is rendered with another. The study for this classification of ragas has been conducted in this paper by implementing a K-Nearest Neighbor (KNN) Classifier, Convolutional Neural Network (CNN) model, and also an ensemble of CNN models. The results of these classifiers are compared using the accuracy metrics, and it was seen that CNN results in 70% accuracy, thereby outperforming KNN with 60% accuracy for the classification of ragas. Furthermore, an ensemble of CNN models gave an accuracy of 60% while having a fast running time.