Efficient Sequence Regression by Learning Linear Models in All-Subsequence Space
Refereed Conference Meeting Proceeding
We present a new approach for learning a sequence regression function, i.e., a mapping from sequential observations to a numeric score. Our learning algorithm employs coordinate gradient descent and Gauss-Southwell optimization in the feature space of all subsequences. We give a tight upper bound for the coordinate wise gradients of squared error loss that enables ecient Gauss-Southwell selection. The proposed bound is built by separating the positive and the negative gradients of the loss function and exploits the structure of the feature space. Extensive experiments on simulated as well as real-world sequence regression benchmarks show that the bound is eective and our proposed learning algorithm is ecient and accurate. The resulting linear regression model provides the user with a list of the most predictive features selected during the learning stage, adding to the interpretability of the method.
ECML PKDD 2017
Digital Object Identifer (DOI):
Macedonia (the former Yugoslav Republic of)
National University of Ireland, Dublin (UCD)
Open access repository: