Exact Inference for Multi-label Classification using Sparse Graphical Models
نویسندگان
چکیده
This paper describes a parameter estimation method for multi-label classification that does not rely on approximate inference. It is known that multi-label classification involving label correlation features is intractable, because the graphical model for this problem is a complete graph. Our solution is to exploit the sparsity of features, and express a model structure for each object by using a sparse graph. We can thereby apply the junction tree algorithm, allowing for efficient exact inference on sparse graphs. Experiments on three data sets for text categorization demonstrated that our method increases the accuracy for text categorization with a reasonable cost.
منابع مشابه
Parameter Learning for Loopy Markov Random Fields with Structural Support Vector Machines
Discriminative learners for functions with structured outputs have been successfully applied to sequence prediction, parsing, sequence alignment, and other problems where exact inference is tractable. However, these learners face challenges when inference procedures must use approximations, e.g., for multi-label classification, clustering, image segmentation, or loopy graphical models. In this ...
متن کاملProbabilistic Matrix Addition
We introduce Probabilistic Matrix Addition (PMA) for modeling real-valued data matrices by simultaneously capturing covariance structure among rows and among columns. PMA additively combines two latent matrices drawn from two Gaussian Processes respectively over rows and columns. The resulting joint distribution over the observed matrix does not factorize over entries, rows, or columns, and can...
متن کاملLarge Margin Boltzmann Machines and Large Margin Sigmoid Belief Networks
Current statistical models for structured prediction make simplifying assumptions about the underlying output graph structure, such as assuming a low-order Markov chain, because exact inference becomes intractable as the tree-width of the underlying graph increases. Approximate inference algorithms, on the other hand, force one to trade off representational power with computational efficiency. ...
متن کاملMulti-Label Classification with Cutset Networks
In this work, we tackle the problem of Multi-Label Classification (MLC) by using Cutset Networks (CNets), weighted probabilistic model trees, recently proposed as tractable probabilistic models for discrete distributions. We employ CNets to perform Most Probable Explanation (MPE) inference exactly and efficiently and we improve a state-of-the-art structure learning algorithm for CNets by explic...
متن کاملMulti-Label Classification Using Conditional Dependency Networks
In this paper, we tackle the challenges of multilabel classification by developing a general conditional dependency network model. The proposed model is a cyclic directed graphical model, which provides an intuitive representation for the dependencies among multiple label variables, and a well integrated framework for efficient model training using binary classifiers and label predictions using...
متن کامل