The present work deals with the automatic detection of the sleep stages from the single-channel EEG data. Various stages of sleep are Awake, sleep stage 1, 2, 3 and 4 and rapid eye movement. Statistical attributes are extracted with the help of Ensemble Empirical Mode Decomposition, Hjorth parameter and zero-crossing rate. Ten-cross fold classification process is followed after best ranked attribute selection. After attributes are selected, the data is classified using bagging classifier. Accuracies of 98.46%, 95.62%, 93.87%, 93.17% and 91.93% for two-stages, three-stages, four-stages, five-stages and six-stages classification respectively. This classifier can be used for the real life application due to higher accuracies.