Identification of Objectionable Audio Segments Based on Pseudo and Heterogeneous Mixture Models.
Ziqiang ShiJiqing HanTieran ZhengJi LiPublished in: IEEE Trans. Speech Audio Process. (2013)
Keyphrases
- mixture model
- em algorithm
- gaussian mixture model
- density estimation
- probabilistic model
- maximum likelihood
- generative model
- multimedia
- unsupervised learning
- expectation maximization
- finite mixtures
- automatic model selection
- model selection
- mixture modeling
- text classification
- semi supervised
- active learning
- data sets
- model based clustering
- finite mixture models