The aim is to identify the actions carried out by a human, from the data collected from the sensors and the surrounding environment. The movement of a human can be recorded through sensors with the help of which activities and events can be recognized. Automatic recognition of activities and events is possible by processing this sensor data with appropriate machine learning and data mining approaches. We will examine several new machine learning and data mining approaches based on decision trees and ensemble learning techniques including random forests and random committee, and compare them with traditional naive Bayes classifier and K-Means clustering approaches for processing sensor signals for activity recognition.