Decision-Making and Learning - Comparing Orthogonal Methods to Majority- Voting

摘要

A study on learning and decision-making methods was conducted by comparing an orthogonal methodology of manipulating data versus that of a majority voting procedure. The latter method has recently become popular in the literature involving applications such as pattern recognition. To evaluate the differences between the proposed methods, data from a multidimensional paradigm involving decision making and learning are analyzed. A number of basic concepts from estimation and information theory are first discussed to understand both the motivation and the underlining issues involved in conducting this study.