This paper proposes a Mutual Information Independence Model (MUM) to segment and label sequential data. MIIM overcomes the strong context independent assumption in traditional generative HMMs by assuming a novel pairwise mutual information independence. As a result, MIIM separately models the long state dependence in its state transition model in a generative way and the observation dependence in its output model in a discriminative way. In addition, a variable-length pairwise mutual information-based modeling approach and a kNN algorithm using kernel density estimation are proposed to capture the long state dependence and the observation dependence respectively. The evaluation on shallow parsing shows that MIIM can effectively capture the long context dependence to segment and label sequential data. It is interesting to note that using kernel density estimation leads to increased performance over using a classifier-based approach. © Springer-Verlag Berlin Heidelberg 2005.
CITATION STYLE
Zhou, G., Yang, L., Su, J., & Ji, D. (2005). Mutual information independence model using kernel density estimation for segmenting and labeling sequential data. In Lecture Notes in Computer Science (Vol. 3406, pp. 155–166). Springer Verlag. https://doi.org/10.1007/978-3-540-30586-6_15
Mendeley helps you to discover research relevant for your work.