Classification model for multi-sensor data fusion apply for Human Activity Recognition

Human Activity Recognition (HAR) have been developed for recognize context generated from human. In real environment system required multi -sensor for support large area and get more accuracy result. Using multi-sensor make high dimensional data which inefficacious for classification model. This paper is concerned with developing classification model that supports high dimensional data and reducing system process by used only some collected data in the decision process. A new-developed model was not developed to be the most accurate but developed to adjust level of credibility. This model was tested using simulated data from real behavior context. Test Results compared with Neural networks (NN) was similar. But developed model uses less data.