| dc.description.abstract | This research has developed an Early Warning System (EWS) integrated with dynamic system simulation and intelligent computation using Artificial Neural Network (ANN) to detect the level of food crisis. The system has been tested and validated using a set of data comprising 28 provinces and 265 districts (kabupaten). The data used for training consits of 167 elements, and the remaing data is used for testing and validation. The accuracy of the sistem to detect the level of food crisis is 96.9%, with mean square error (MSE) equal to 0.11. Food crisis factors and parameters together with variables derived from the identified parameters have been formulated from testing and validation of the system prototype and the analysis of the system output of ANN. It can the be identified that the weight priority of all variables are shown in decreasing order with respect to weight as follows: 1). Natural Disaster (X5), 2). Pepople under poverty line (X4), 3). Infant mortality (X3), 4). IHSG (X10), 5). Infant underweight (X2), 6). Price of rice (X8), 7). Area without forest (X6), 8). Normative Consumption Ratio (XI), 9). Annual Rainfall (X7), and 10). Dollars Exchange (X9). Factor interactios that relate to food food vulnerability is complex, dynamic, and probalistic involving multi aspects and multi dimensions. Dynamic system simulation unified with an intelligent computation using Artificial Neural Network (ANN) can be utilized to cope with criticallity of such factor interactions that influence food crisis. | en |
