Transductive Conndence Machines for Pattern Recognition
نویسندگان
چکیده
We propose a new algorithm for pattern recognition that outputs some measures of \reliability" for every prediction made, in contrast to the current algorithms that output \bare" predictions only. Our method uses a rule similar to that of nearest neighbours to infer predictions ; thus its predictive performance is close to that of nearest neighbours, while the measures of conndence it outputs provide practically useful information for individual predictions.
منابع مشابه
Computationally E cient Transductive Machines
In this paper we propose a new algorithm for providing conndence and credibility values for predictions on a multi-class pattern recognition problem which uses Support Vector machines in its implementation. Previous algorithms which have been proposed to achieve this are very processing intensive and are only practical for small data sets. We present here a method which overcomes these limitati...
متن کاملInductive Confidence Machines for Regression
The existing methods of predicting with conndence give good accuracy and conndence values, but quite often are computationally inee-cient. Some partial solutions have been suggested in the past. Both the original method and these solutions were based on transductive inference. In this paper we make a radical step of replacing transductive inference with inductive inference and deene what we cal...
متن کاملTransductive Confidence Machines for Pattern Recognition
We propose a new algorithm for pattern recognition that outputs some measures of \reliability" for every prediction made, in contrast to the current algorithms that output \bare" predictions only. Our algorithm is based on the well-known nearest neighbours algorithm and uses a similar rule to infer predictions. Introduction 1
متن کاملMachine-Learning Applications of Algorithmic Randomness
Most machine learning algorithms share the following drawback: they only output bare predictions but not the conndence in those predictions. In the 1960s algorithmic information theory supplied universal measures of conndence but these are, unfortunately, non-computable. In this paper we combine the ideas of algorithmic information theory with the theory of Support Vector machines to obtain pra...
متن کاملGlobal, local and personalised modeling and pattern discovery in bioinformatics: An integrated approach
The paper is a comparative study of major modeling and pattern discovery approaches applicable to the area of Bioinformatics and the area of decision support systems in general. These approaches include inductive versus transductive reasoning, global, local, and personalized modeling and their potentials are illustrated on a case study of gene expression and clinical data related to cancer outc...
متن کامل