Majority vote is a much studied topic; in particular, the well-known Condorcet Jury Theorem (CJT) had provided validity to the belief that the majority opinion of a group is superior to those of individuals provided the individuals have reasonable competence. The mathematics of the theorem can be easily explained to teachers with a basic knowledge of statistics, and its interpretation has considerable social implications. Recently, the voting principle has been applied to the technological area of optical character recognition, with significant results. In this domain, computer programs are designed and implemented to read or identify characters and words. When the data is handwritten, the immense diversity in writing styles makes it extremely difficult for one program operating alone to achieve the high levels of accuracy that are required for practical applications. On the other hand, many different programs have been designed for this purpose. Consequently, researchers began to combine the results of different programs by majority vote, in order to obtain more accurate performance. This has created a new trend in pattern recognition, and in the process this author has also derived new results about majority vote. These theoretical findings are reflected in experimental results, and the process has provided an example of the interaction between basic mathematical ideas and their applications in advanced technology.
|Publication status||Published - Nov 2000|