Research on garment image classification algorithm based on machine learning

With the development of e-commerce economy, more and more people prefer online shopping. Therefore, if we can accurately detect clothing images, it is of great significance in enhancing users' shopping experience and improving e-commerce service quality. Based on machine learning technology, this paper uses traditional machine learning algorithms such as random forest algorithm, k-nearest neighbor classification algorithm and naive Bayes to establish clothing image classification model, and uses deep learning to build convolution neural networks with four different optimizers: Momentum, RMSprop, Adam and SGD. By optimizing each algorithm model, the classification accuracy is improved on the fashionmnist dataset. Experiments show that compared with the traditional machine learning algorithm, the deep learning algorithm has higher prediction accuracy in this paper for image classification problems, and in deep learning, the deep learning algorithm using SGD optimizer can achieve higher accuracy and better prediction effect.