Comparing k-Nearest Neighbors and Convolutional Neural Networks for Digit Recognition: A Performance and Computational Complexity Analysis
Keywords:
digit recognition, Big-Data, MNIST dataset, k-Nearest Neighbors, Convolutional Neural Networks, machine learning, computer vision, performance evaluation, computational complexityAbstract
Digit recognition is a fundamental task in computer vision, and various algorithms have been developed for this purpose. In this study, we compare the performance and computational complexity of two popular algorithms, k-Nearest Neighbors (kNN) and Convolutional Neural Networks (CNN), for digit recognition on the widely used MNIST dataset. We evaluate the algorithms based on their accuracy, F1 score, training time, and evaluation time. Our results show that the CNN algorithm outperforms the kNN algorithm in terms of accuracy and F1 score, with a statistically significant difference between the algorithms. The trade-off between computational complexity and performance is also highlighted, with the kNN algorithm having a relatively low training time but a high evaluation time, and the CNN algorithm having a longer training time but a much faster evaluation time. Our study provides insights into the performance and computational complexity of these algorithms and can inform the selection of algorithms for similar image classification tasks.
References
Goodfellow, I., Bengio, Y., and Courville, A. (2016). Deep learning. MIT Press.
Hastie, T., Tibshirani, R., and Friedman, J. (2017). The elements of statistical learning: Data mining, inference, and prediction. Springer.
Alpaydin, E. (2010). Introduction to machine learning. MIT Press.
Bishop, C. M. (2006). Pattern recognition and machine learning. Springer.
