Analysis of Weather and Exercise Steps Based on LSTM Neural Network

The analysis of exercise and health has become a hot issue that people are paying attention to today. This paper presents an automated method for the complete analysis of weather and air quality and daily exercise volume (number of steps). Firstly, a clustering method based on Fourier transform is used to separate users of different exercise styles, and then the LSTM neural network is used to train the step data and the corresponding weather and air quality data to find the correlation. After testing the real data with multiple exercise styles, the validity of this analysis method can be confirmed. At the same time, the training verification time of the model is also collected, which proves that it has certain advantages in the analysis of large-scale data.