Time Series Classification using Attention-Based LSTM and CNN
R Vaibhava Lakshmi, S. Radha · 2023
Time series classification has many real-world applications, including medical diagnosis, financial forecasting, and environmental monitoring. Accurately classifying time series data can provide valuable insights and help make informed decisions in various fields. This paper proposes the attention-based LSTM - CNN framework for classifying time series data. Our framework incorporates joint extraction of spatio-temporal features from a convolutional neural network and attention-based long- and short-term memory. The extracted features are fed into a Random Forest classifier to obtain the final output classes. We have used the Wafer dataset from the UCR repository for experimentation. An accuracy of 99.86% was achieved for the proposed model, hence transcending the performance of the other baseline models.