Emotion recognition plays a crucial role in the era of artificial intelligence and the internet of things. It offers tremendous scope to human-computer interaction, robotics, health care, biometric security, and behavioral modeling. Emotion recognition systems recognize emotions from facial expressions, text data, body movements, voice, brain, or heart signals. Along with basic emotions, attitude, control over emotions, and power of activation of emotion can also be examined for analyzing sentiments. This paper identifies variously supervised and unsupervised machine-learning techniques for feature extraction and emotion classification. Comparative analysis has also been made of various machine-learning algorithms used in referenced papers. It tells the scope and applications of automatic emotion recognition systems in various fields. This paper also discusses various parameters to increase the accuracy, security, and efficiency of the system.