An information theoretic framework for weight estimation in the combination of probabilistic classifiers for speaker identification

Altincay, H
Demirekler, Mübeccel
In this paper, we describe a relation between classification systems and information transmission systems. By looking at the classification systems from this perspective, we propose a method of classifier weight estimation for the linear (LIN-OP) and logarithmic opinion pool (LOG-OP) type classifier combination schemes for which some tools from information theory are used. These weights provide contextual information about the classifiers such as class dependent classifier reliability and global classifier reliability. A measure for decision consensus among the classifiers is also proposed which is formulated as a multiplicative part of the classifier weights. A method of selecting the classifiers which provide complementary information for the combination operation is given. Using the proposed method, two classifiers are selected to be used in the combination operation. Simulation experiments in closed set speaker identification have shown that the method of weight estimation described in this paper improved the identification rates of both linear and logarithmic opinion type combination schemes. A comparison between the proposed method and some other methods of weight selection is also given at the end of the paper.


A statistical unified framework for rank-based multiple classifier decision combination
Saranlı, Afşar (2001-04-01)
This study presents a theoretical investigation of the rank-based multiple classifier decision combination problem, with the aim of providing a unified framework to understand a variety of such systems. The combination of the decisions of more than one classifiers with the aim of improving overall system performance is a concept of general interest in pattern recognition, as a viable alternative to designing a single sophisticated classifier. The problem of combining the classifier decisions in the raw form...
The principles of B-smooth discontinuous flows
Akalin, E; Akhmet, Marat (Elsevier BV, 2005-05-01)
In this paper, we define B-smooth discontinuous dynamical systems which can be used as models of various processes in mechanics, electronics, biology, and medicine. We find sufficient conditions to guarantee the existence of such systems. These conditions are easy to verify. Appropriate examples are constructed.
A new structural modification method with additional degrees of freedom for dynamic analysis of large systems
Şayin, Burcu; Ciğeroğlu, Ender (2013-02-14)
Structural modification techniques are widely used to predict the dynamic characteristics of modified systems by using the information of the original and modifying systems. Nowadays, due to the increased computational power, large finite element models are regularly used. During design optimization, modifications are done on the original structure in order to meet the design requirements which require reevaluation of the dynamic response of the structure. This is a time consuming process since for each mod...
A finite field framework for modeling, analysis and control of finite state automata
Reger, Johann; Schmidt, Klaus Verner (Informa UK Limited, 2004-09-01)
In this paper, we address the modeling, analysis and control of finite state automata, which represent a standard class of discrete event systems. As opposed to graph theoretical methods, we consider an algebraic framework that resides on the finite field F-2 which is defined on a set of two elements with the operations addition and multiplication, both carried out modulo 2. The key characteristic of the model is its functional completeness in the sense that it is capable of describing most of the finite st...
A temporal neurofuzzy model for rule-based systems
Alpaslan, Ferda Nur; Jain, L (1997-05-23)
This paper reports the development of a temporal neuro-fuzzy model using fuzzy reasoning which is capable of representing the temporal information. The system is implemented as a feedforward multilayer neural network. The learning algorithm is a modification of the backpropagation algorithm. The system is aimed to be used in medical diagnosis systems.
Citation Formats
H. Altincay and M. Demirekler, “An information theoretic framework for weight estimation in the combination of probabilistic classifiers for speaker identification,” SPEECH COMMUNICATION, pp. 255–272, 2000, Accessed: 00, 2020. [Online]. Available: