Codes for window ranking and EEG feature extraction (window selection criteria based and time segmentation based)

Electroencephalograph (EEG)-based emotion recognition has become increasinglypopular in the field of Brain-Computer Interface (BCI) due to the non-invasive natureand high temporal resolution of EEG. However, previous studies have primarilyfocused on analyzing the entire EEG data segment for the EEG-emotion recognitiontask, even though emotions are highly localized in time and of limited duration.Therefore, considering a shorter time frame from EEG data for emotion recognitionwould be a more reasonable alternative. One of the key indicators of emotion in EEGsignals is the various EEG frequency bands. Hence, in this paper, we propose a novelapproach for localizing temporal windows from emotional EEG data based on EEGfrequency band power and analyze their influence in temporal window localization forefficiency enhancement of EEG-based emotion recognition. In addition, we present aframework for feature extraction based on time segmentation, which is applied totemporally localized EEG data segments. We furthermore evaluated the efficacy ofseveral conventional feature extraction methods within the EEG-based emotionrecognition framework. The proposed methodologies have been thoroughly evaluatedon three publicly accessible datasets, DEAP, SEED-IV, and SEED, using threedifferent classification approaches, and the experimental results demonstrated superiorperformance compared to several state-of-the-art methods.