Abstract
With the development of artificial intelligence and the improvement of hardware computing power, deep learning models have become widely used in the Internet of Things (IoT) field, especially for analyzing spatiotemporal data collected by wireless sensors. Recurrent Neural Networks (RNN) such as Long Short-Term Memory (LSTM) network are generally used for these time series data. Hyperparameter settings of model training are regarded as essential factors for the performance of deep learning models. Manually optimizing hyperparameters not only cost more resources but also be more likely to set hyperparameters that follow stereotypes, resulting in unreasonable hyperparameter settings and poor model performance. As one of the most important fields in Automated Machine Learning (AutoML) research, automated hyper-parameter optimization (HPO) mainly includes grid search, hyperparameter search based on genetic algorithm, etc. Whereas these methods have their own drawbacks. In this paper, an automated hyper-parameter optimization method based on parallel genetic algorithm (PGA) is proposed. According to the process of PGA, this paper divided hyperparameter optimization into several stages, including population initialization, fitness function, tournament selection, crossover operators, mutation operators, subgroup exchange, and end of evolution. Then the proposed hyperparameter optimization method is implemented in LSTM models and tested on two different time series datasets collected by real-world IoT sensors. By comparing our proposed method with other mainstream HPO methods in different datasets, it is proved that our HPO method based on PGA shows better performance on both time costs and prediction results.
Funding Information
  • National Natural Science Foundation of China (61972145)
  • National Key R&D Program of China (2019YFB1405703)

This publication has 22 references indexed in Scilit: