 Independent component analysis, ICA, is a widely used technique for removing noise from multi-channel electroencephalography, EEG, signals. However, it can be time-consuming and requires expertise which is often unavailable. In order to overcome this limitation, research has been conducted to develop automated models for identifying independent components, ICS. These models typically rely on power spectrum density, PSD, and top-a-plot data, but do not take into account the time series data which could provide additional insight into the signal. Our proposed method uses all three sources of data to improve IC classification accuracy. Experiments show that incorporating time series data significantly improved the classification accuracy, compared to existing methods. Furthermore, transfer learning was found to be more effective than training a new model from scratch. This work demonstrates the potential of combining multiple sources of data to improve IC classification accuracy, and suggests that transfer learning should be considered when developing new deep learning models. This article was authored by Fabio Lopez, Adriana Leo, Julio Maderos, and others. We are article.tv, links in the description below.