TOWARDS GROUP-AWARE LEARNING ANALYTICS: USING SOCIAL NETWORK ANALYSIS AND MACHINE LEARNING TO MONITOR AND PREDICT PERFORMANCE IN COLLABORATIVE LEARNING