Date on Master's Thesis/Doctoral Dissertation
Bioinformatics and Biostatistics
An ensemble consists of a set of individual predictors whose predictions are combined. Generally, different classification and regression models tend to work well for different types of data and also, it is usually not know which algorithm will be optimal in any given application. In this thesis an ensemble regression function is presented which is adapted from Datta et al. 2010. The ensemble function is constructed by combining bagging and rank aggregation that is capable of changing its performance depending on the type of data that is being used. In the classification approach, the results can be optimized with respect to performance measures such as accuracy, sensitivity, specificity and area under the curve (AUC) whereas in the regression approach, it can be optimized with respect to measures such as mean square error and mean absolute error. The ensemble classifier and ensemble regressor performs at the level of the best individual classifier or regression model. For complex high-dimensional datasets, it may be advisable to combine a number of classification algorithms or regression algorithms rather than using one specific algorithm.
Shah, Jasmit SureshKumar, "An adaptive ensemble learner function via bagging and rank aggregation with applications to high dimensional data." (2011). Electronic Theses and Dissertations. Paper 1306.