Efficient EM Training of Gaussian Mixtures with Missing Data
This work addresses a common issue in data mining where missing data hinders machine learning, offering an incremental improvement for practitioners dealing with incomplete datasets.
The paper tackles the problem of training Gaussian mixture models with missing data by introducing a spanning-tree algorithm that speeds up computation, achieving significant efficiency gains in data-mining applications.
In data-mining applications, we are frequently faced with a large fraction of missing entries in the data matrix, which is problematic for most discriminant machine learning algorithms. A solution that we explore in this paper is the use of a generative model (a mixture of Gaussians) to compute the conditional expectation of the missing variables given the observed variables. Since training a Gaussian mixture with many different patterns of missing values can be computationally very expensive, we introduce a spanning-tree based algorithm that significantly speeds up training in these conditions. We also observe that good results can be obtained by using the generative model to fill-in the missing values for a separate discriminant learning algorithm.