Proceedings of the ... SIAM International Conference on Data Mining. SIAM International Conference on Data Mining最新文献

HIV rapidly evolves drug resistance in response to antiviral drugs used in AIDS therapy. Estimating the specific resistance of a given strain of HIV to individual drugs from sequence data has important benefits for both the therapy of individual patients and the development of novel drugs. We have developed an accurate classification method based on the sparse representation theory, and demonstrate that this method is highly effective with HIV-1 protease. The protease structure is represented using our newly proposed encoding method based on Delaunay triangulation, and combined with the mutated amino acid sequences of known drug-resistant strains to train a machine-learning algorithm both for classification and regression of drug-resistant mutations. An overall cross-validated classification accuracy of 97% is obtained when trained on a publically available data base of approximately 1.5×10⁴ known sequences (Stanford HIV database http://hivdb.stanford.edu/cgi-bin/GenoPhenoDS.cgi). Resistance to four FDA approved drugs is computed and comparisons with other algorithms demonstrate that our method shows significant improvements in classification accuracy.

The focus of this paper is on how to select a small sample of examples for labeling that can help us to evaluate many different classification models unknown at the time of sampling. We are particularly interested in studying the sampling strategies for problems in which the prevalence of the two classes is highly biased toward one of the classes. The evaluation measures of interest we want to estimate as accurately as possible are those obtained from the contingency table. We provide a careful theoretical analysis on sensitivity, specificity, and precision and show how sampling strategies should be adapted to the rate of skewness in data in order to effectively compute the three aforementioned evaluation measures.

State-of-the-art learning algorithms accept data in feature vector format as input. Examples belonging to different classes may not always be easy to separate in the original feature space. One may ask: can transformation of existing features into new space reveal significant discriminative information not obvious in the original space? Since there can be infinite number of ways to extend features, it is impractical to first enumerate and then perform feature selection. Second, evaluation of discriminative power on the complete dataset is not always optimal. This is because features highly discriminative on subset of examples may not necessarily be significant when evaluated on the entire dataset. Third, feature construction ought to be automated and general, such that, it doesn't require domain knowledge and its improved accuracy maintains over a large number of classification algorithms. In this paper, we propose a framework to address these problems through the following steps: (1) divide-conquer to avoid exhaustive enumeration; (2) local feature construction and evaluation within subspaces of examples where local error is still high and constructed features thus far still do not predict well; (3) weighting rules based search that is domain knowledge free and has provable performance guarantee. Empirical studies indicate that significant improvement (as much as 9% in accuracy and 28% in AUC) is achieved using the newly constructed features over a variety of inductive learners evaluated against a number of balanced, skewed and high-dimensional datasets. Software and datasets are available from the authors.