Journal of Applied Sciences1812-56541812-5662Asian Network for Scientific Information10.3923/jas.2013.1728.1733ZhaoXiukuan NingBaiqi SongGangbing 1020131310Conventional artificial intelligent methods such as neural network and SVM use only labeled data (feature/label pairs) for training. Labeled instances are often difficult, expensive, or time consuming to obtain. To use a large amount of unlabeled data together with labeled data to build better models, we proposed an active semi-supervised learning method based on the pool query active learning and manifold regularization semi-supervised method. In this paper, the effectiveness of the method was verified by its application to a synthetic data set and three real world classification problems. The experimental results showed that employing our active semi-supervised learning method can significantly reduce the need for labeled training instances.]]>Bache, K. and M. Lichman,20132013Belkin, M., P. Niyogi and V. Sindhwani,20061223992434Chapelle, O., V. Sindhwani and S.S. Keerthi,20089203233Liu, Y.,20044419361941Saffari, A., C. Leistner and H. Bischof,20092009Settles, B.,20102010Shen, J., B. Ju, T. Jiang, J. Ren, M. Zheng, C. Yao and L. Li,20117437853792Tong, S. and D. Koller,200124566Weston, J., F. Ratle, H. Mobahi and R. Collobert,20122012pp: 639-655pp: 639-655Wu, J., Y.B. Diao, M.L. Li, Y.P. Fang and D.C. Ma,20091151155Yu, D., B. Varadarajan, L. Deng and A. Acero,201024433444Zhang, C. and T. Chen,20024260268Zhao, X., M. Li, J. Xu and G. Song,2011381019910204Zhao, X., X. Li, C. Pang and S. Wang,20131054550Zha, Z.J., M. Wang, Y.T. Zheng, Y. Yang, R. Hong and T.S. Chua,2012141727Zhu, X.J.,20082008Zhu, X., J. Lafferty and Z. Ghahramani,20032003pp: 5865