4.4. Data Loading and Parsing

Figure 5: DREAMER Dataset Structure in MATLAB

Figure 6: Data Ingestion and Parsing Workflow for the DREAMER Dataset

Figure 5 shows the data structure in MATLAB and Figure 6 presents overview of the process. The data loading and parsing process, as shown in Figure 6, begins by importing essential libraries to handle and visualize the data efficiently. The DREAMER dataset is loaded using scipy.io and converted into a JSON structure to facilitate easier data extraction. A dedicated function then extracts key participant emotion data, organizing it into a matrix that captures demographic details, video information, and emotion ratings. This matrix is reshaped into a DataFrame, which consolidates participant age, gender, video names, target emotions, and valence, arousal, and dominance scores. Additionally, summary statistics, including mean and standard deviation for each video, are computed and saved for subsequent analysis.