An Approach to Effective Processing of Information from Smart Home Sensors using Machine Learning Methods

The article analyzes various machine learning algorithms that can be used to process information from sensors in a smart home system. Based on the results of the analysis, a new approach to efficient processing of information received from smart home sensors is proposed, based on the use of the MobileNet Convolutional Neural Network architecture and the TensorFlow machine learning software library, as well as the TensorBoard monitoring tool. Experiments have shown that the proposed approach allows achieving a sufficiently high accuracy and efficiency of information processing.