An introduction to variable and feature selection
The Journal of Machine Learning Research
Pattern Recognition and Machine Learning (Information Science and Statistics)
Pattern Recognition and Machine Learning (Information Science and Statistics)
Learning Spectral Clustering, With Application To Speech Separation
The Journal of Machine Learning Research
CARA: A Cultural-Reasoning Architecture
IEEE Intelligent Systems
Social Computing: From Social Informatics to Social Intelligence
IEEE Intelligent Systems
Guest Editors' Introduction: Social Computing
IEEE Intelligent Systems
Finding Most Probable Worlds of Probabilistic Logic Programs
SUM '07 Proceedings of the 1st international conference on Scalable Uncertainty Management
CONVEX: Similarity-Based Algorithms for Forecasting Group Behavior
IEEE Intelligent Systems
IEEE Intelligent Systems
Fast approximate spectral clustering
Proceedings of the 15th ACM SIGKDD international conference on Knowledge discovery and data mining
Stepwise nearest neighbor discriminant analysis
IJCAI'05 Proceedings of the 19th international joint conference on Artificial intelligence
Multilevel manifold learning with application to spectral clustering
CIKM '10 Proceedings of the 19th ACM international conference on Information and knowledge management
| Hi-index | 0.00 |
Researchers have developed methods to predict a terrorist organization's probable actions (such as bombings or kidnappings). The terrorist organization's actions can be affected by the context of the organization. Thus, the organization's context variables can be used to improve the accuracy of forecasting the terrorist behavior. Those algorithms based on context similarity suffer a serious drawback that it can result in the algorithm's fluctuation and reduce the prediction accuracy if not all the attributes are detected. A prediction algorithm PBCS (Prediction Based on Context Subspace) based on context subspace is proposed in this paper. The proposed algorithm first extracts the context subspace according to the association between the context attributes and the behavior attributes. Then, it predicts the terrorist behavior based on the extracted context subspace. The proposed algorithm uses the improved spectral clustering method to obtain the context subspace. It concerns the distribution of the data samples, the label information and the local similarity of the data in the process of extracting the subspace. Experimental results on the artificial dataset and the MAROB dataset show that the prediction method proposed in this paper can not only improve the prediction accuracy but also reduce the prediction fluctuation.