Cost-Sensitive Analysis in Multiple Time Series Prediction

In this paper we propose a new methodology for Cost-Benefit analysis in a multiple time series prediction problem. The proposed model is evaluated in a real world application based on a network of wireless sensors distributed in energy production plants in a region. These sensors generate multiple time series data representing energy production. To build the prediction model for total energy production in the region we have used three common forecasting techniques, Support Vector Machines (SVMs), Multilayer Perceptron (MLP), and Multiple Regression (MR). For training and testing of the models we have used the data from year 2002 to 2004. We analyzed the quality of total energy prediction with different subsets of sensors. We build our cost-benefit model for the prediction process as a function of sensors in a distributed network and estimated the optimum number of sensors that will balance the expenses of the system with the prediction accuracy.