An improved BP neural network in Internet of Things data classification application research

Abstract

For the problem that traditional BP neural network exists slow convergence, easy to fall into local minimum value which will restrict the application of BP neural network in the Internet of Things(IOT), this paper proposes an improved BP algorithm which integrates into variable learning rate. In order to verify the effectiveness of the improved algorithm, the paper makes use of neural networks toolkit and GUI interface to classify collected data experiment based on MATLAB environment. Experimental result shows that the algorithm has better data classification performance.