Modified MC-MHLF Model using Multi-Channel Input for Time Series Classification

Classification of time series data has gathered significant interest in various fields owing to the wide application range of predicting classification labels for unclassified time series data. Specifically, the high-precision classification of time series data has numerous potential applications in the real world, such as the classification of sensor measurement data. Therefore, studies have been conducted on time series classification using machine learning to achieve high accuracy. This study aims to improve the accuracy of classifying time series data with enhancing multi-channel MHLF (MC-MHLF) model. The MC-MHLF model is based on the long short term memory fully convolutional network (LSTM-FCN) model with multi-channel input data. In this study, we modify the existing model, adding the moving average convergence divergence (MACD) to the multi-channel data provided as input to the MC-MHLF model. Using the time series classification archive of the UCR, the evaluation experiments show that the proposed model is more accurate than the previous model. In addition, the proposed model is more accurate than existing time series classification models based on machine learning.