Artificial Neural Network: Approach for Classifying Trajectories on Road Network
Deepak S. Gaikwad, Usha A. Jogalekar · 2014
Classification is the classical problem in the process of machine learning and data mining. Classification is widely used for modeling various types of data sets, namely sets of items, text document, networks, signals and graphs. But, there is lack of study on Trajectory Classification. Trajectory classification is defined as the process of predicting vehicles class label based on its trajectory. Trajectory Data is the data which is collected from moving objects (vehicles). Large amount of trajectory (i.e. vehicles on road network) data is collected from technologies such as GPS, Sensors, Cameras, RFID etc., this data is useful for better Traffic prediction, activity recognition as well as transportation planning of the city. In this paper, Multi-Layer framework of Artificial Neural Network (ANN) is applying for classifying the trajectories on road network. ANN’s are generally presented as systems of interconnected neurons which compute values from inputs. ANN is used for classification of items, signals etc. ANN is also employed in analysis of investment, stock exchange as well as verification of signature etc., but less employed on Trajectory data. In this paper, analyzing behavior of trajectories on road network, we have employed ANN on Trajectory data and projecting a model using sequential pattern. Sequential patterns are good feature candidate as it preserves order of visiting sequence of trajectories on road network.