LSB_CMM is the Matlab implementation of Conditional Multinomial Mixture Model for Superset Label Learning(L. Liu et al.). This method is based on Logistic Breaking process and EM algorithm. lsbcmm_fit.m is the main function. Given the parameters, it will fit the model for this partial label learning problem. runExp.m runs experiments on UCI datasets and three real-world datasets.
Reference: [1] Jin R, Ghahramani Z. Learning with multiple labels. In: Advances in Neural Information Processing Systems 15, Cambridge, MA: MIT Press, 2003, 921-928. [2] Liu L, Dietterich T. A conditional multinomial mixture model for superset label learning. In: Advances in Neural Information Processing Systems 25, Cambridge, MA: MIT Press, 2012, 557-565. [3] 张敏灵. 偏标记学习研究综述. 数据采集与处理, 2015, 30(1): 77-87.