Done as a part of CSE 572 Data Mining Course work in MATLAB In this project, firstly an attempt was made to understand human activities that can identify eating with spoon or fork or not eating with them. The real-world Mayo data consists of Wristband Sensor, IMU and EMG sensors, but in this project only one sensor data is used for all users i.e EMG sensor which has a sampling rate of 100 Hz. For identification of activities, initially the data is prepossessed, and useful features are extracted. After that Principal Component Analysis (PCA) is applied understand which features are the dominant ones. Secondly, human activities are understood based on eating or non-eating activities. Also, user dependent analysis and user independent analysis is done across all the 30 users containing in the data set using three machine learning and their performance metrics have been evaluated based on their respective predictions.