COMPARATIVE ANALYSIS THE PERFORMANCE OF CLIENT-SIDE AND SERVER-SIDE MACHINE LEARNING TECHNOLOGIES
Abstract
The performance analysis of client-side and server-side machine learning technologies is important for understanding the optimal way to model optimization. The study aims to analyze the training time of the model, taking into account parameters such as the number of likes, comments and shares according to the text of a post in social networks. Natural language processing (NLP) requires significant computing power, so it is important to determine whether it is more efficient to train models on client devices or on servers. TensorFlow for JavaScript can provide client-side computation, while Python can use server-side resources. The obtained results confirm that the models in web machine learning require optimization and are slower than in the server implementation, taking into account the training execution time. Therefore, the size of the data is important for effective machine learning of the model in client-side computing. Keywords : machine learning, client-side computing, server-side computing, model training, text analysis, natural language processing, TensorFlow, learning time, social networks.