Simulation of Multidimensional Time Series Data Analysis Model Based on Deep Learning

By analyzing time series, we can realize functions such as prediction and detection to save manpower and material resources. However, time series data are usually accompanied by noise and data loss, which greatly restricts our use and analysis of time series data. In this paper, the current situation of time series classification research is comprehensively analyzed, and a multi-dimensional time series data analysis model based on deep learning is proposed. The feature extraction part of the model consists of a hollow convolution space pyramid structure and two residual blocks, and the residual blocks follow the structure of ResNet classification model. The pyramid structure of empty convolution space can be used as a basic module structure and a part of other types of neural network structures to obtain rich feature information, or it can be simply stacked many times and used as an independent network structure. Experimental results show that the proposed model has similar and good classification performance. Compared with other algorithms, the end-to-end deep learning algorithm designed in this paper has greatly improved the accuracy and solved the problem of the accuracy of multi-dimensional time series classification.