Abstract—Investigating the mental fitness is a key research field in medical domain because of its significance. The state and the activity of brain can be examined based the brain signals and the brain images. The existing brain imaging procedures are costlier than the brain signal evaluating approach. Hence, during the preliminary examination, the condition of the brain is normally evaluated based on its ElectroEncephaloGram (EEG) recorded through dedicated electrodes. In medical discipline, EEG is widely considered to record the Electrical Activity (EA) of brain. After recording the EEG, its pattern is to be examined physically or by means of an automated procedure to categorize the mental health of patient into normal and abnormal condition. The work proposed in this paper considers a benchmark EEG dataset to classify the signal into normal and abnormal. During the experimental investigation, the amplitude and the entropy based features are extracted for the benchmark EEG signals and the classification is implemented using the random forest (RF) classifier. The experimental investigation of this paper confirms that, the amplitude and entropy based evaluation offers better result on the benchmark dataset. 
 

Index Terms—brain state, EEG signal, amplitude level, entropy value, classification