Towards Automated Classifier Combination for Pattern Recognition

1997

In the last years, great attention has been devoted to multiple classifier systems. The implementation of such a system implies the definition of a rule (combining rule) for determining the most likely class, on the basis of the class attributed by each single expert. The availability of a criterion to evaluate the credibility of the decision taken by a classifier can be profitable in order to implement the combining rule. We propose a method that, after defining the reliability of a classification on the basis of information directly derived from the output of the classifier, uses this information in the context of a combining rule. The results obtained by combining four handwritten character recognizers on the basis of classification reliability are compared with those obtained by using three different combining criteria. Tests have been performed using a standard handwritten character database.

2014

The One-Class Classifier (OCC) has been widely used for solving the one-class and multi-class classification problems. Its main advantage for multi-class is offering an open system and therefore allows easily extending new classes without retraining OCCs. However, extending the OCC to the multi-class classification achieves less accuracy comparatively to other multiclass classifiers. Hence, in order to improve the accuracy and keep the offered advantage we propose in this paper a Multiple Classifier System (MCS), which is composed of different types of OCC. Usually, the combination is performed using fixed or trained rules. Generally, the static weighted average is considered as straightforward combination rule. In this paper we propose a dynamic weighted average rule that calculates the appropriate weights for each test sample. Experimental results conducted on several real-world datasets proves the effective use of the proposed multiple classifier system where the dynamic weighted average rule achieves the best results for most datasets versus the mean, max, product and the static weighted average rules.

2012 International Conference on Frontiers in Handwriting Recognition, 2012

Classifier combination methods have proved to be an effective tool for increasing the performance in pattern recognition applications. The rationale of this approach follows from the observation that appropriately diverse classifiers make uncorrelated errors. Unfortunately, this theoretical assumption is not easy to satisfy in practical cases, thus reducing the performance obtainable with any combination strategy. In this paper we propose a new weighted majority vote rule which try to solve this problem by jointly analyzing the responses provided by all the experts, in order to capture their collective behavior when classifying a sample. Our rule associates a weight to each class rather than to each expert and computes such weights by estimating the joint probability distribution of each class with the set of responses provided by all the experts in the combining pool. The probability distribution has been computed by using the naive Bayes probabilistic model. Despite its simplicity, this model has been successfully used in many practical applications, often competing with much more sophisticated techniques. The experimental results, performed by using three standard databases of handwritten digits, confirmed the effectiveness of the proposed method.

2000

In many applications of computer vision, combination of decisions from mulitiple sources is a very important way of achieving more accurate and robust classification. Many such techniques can be used, two of which are the Majority Voting and the Divide and Conquer techniques. The former achieves decision combination by measuring consensus among the participating classifiers and the latter achieves the same by dividing the problem into smaller problems and solving each of these sub-problems more e f iciently. Both these approaches have their advantages and disadvantages. In this paper, a novel approach t o combining these two techniques is presented. Although the success of the approach has been demonstrated in a typical application area of computer vision (recognition of complex and highly variable image data), the approach is completely generalised and is applicable to other task domains.

This paper presents a method for combining classifiers that uses estimates of each individual classifier's local accuracy in small regions of feature space surrounding an unknown test sample. An empirical evaluation using five real data sets confirms the validity of our approach compared to some other Combination of Multiple Classifiers algorithms. We also suggest a methodology for determining the best mix of individual classifiers.

2004

The aim of this paper is to investigate the role of the a-priori knowledge in the process of classifier combination. For this purpose three combination methods are compared which use different levels of a-priori knowledge. The performance of the methods are measured under different working conditions by simulating sets of classifier with different characteristics. For this purpose, a random variable is used to simulate each classifier and an estimator of stochastic correlation is used to measure the agreement among classifiers. The experimental results, which clarify the conditions under which each combination method provides better performance, show to what extend the a-priori knowledge on the characteristics of the set of classifiers can improve the effectiveness of the process of classifier combination.

Information Fusion, 2001

This paper proposes a novel algorithm for multiple classi®ers combination based on clustering and selection technique (called M3CS), which can ®nd in the feature space the regions where each classi®er has best classi®cation performance. The proposed method may be divided into two steps: clustering and selection (operation). At clustering step, the feature space is partitioned into several regions by clustering separately the correctly and incorrectly classi®ed training samples from each classi®er, and the performances of the classi®er in each region are calculated. In the selection step, the most accurate classi®er in the vicinity of the input sample is nominated to provide the ®nal decision of the committee. The performance comparison between M3CS and Kuncheva's CS DT method, as well as some simple aggregation methods such as maximum, minimum, average, and majority vote, con®rms the validity of the proposed scheme.

2002

A new approach to serial multi-stage combination of classifiers is proposed. Each classifier in the sequence uses a smaller subset of features than the subsequent classifier. The classification provided by a classifier is rejected only if its decision is below a pre-defined confidence level. The approach is tested on a two-stage combination of k-Nearest Neighbor classifiers. The features to be used by the first classifier in the combination are selected by two stand-alone algorithms (Relief and Info-Fuzzy Network, or IFN) and a hybrid method, called "IFN + Relief." The feature-based approach is shown empirically to provide a substantial decrease in the computational complexity, while maintaining the accuracy level of a single-stage classifier or even improving it.

Neurocomputing, 2018

Classifier ensembles are pattern recognition structures composed of a set of classification algorithms (members), organized in a parallel way, and a combination method with the aim of increasing the classification accuracy of a classification system. In this study, we investigate the application of a generalized mixture (GM) functions as a new approach for providing an efficient combination procedure for these systems through the use of dynamic weights in the combination process. Therefore, we present three GM functions to be applied as a combination method. The main advantage of these functions is that they can define dynamic weights at the member outputs, making the combination process more efficient. In order to evaluate the feasibility of the proposed approach, an empirical analysis is conducted, applying classifier ensembles to 25 different classification data sets. In this analysis, we compare the use of the proposed approaches to ensembles using traditional combination methods as well as the state-of-the-art ensemble methods. Our findings indicated gains in terms of performance when comparing the proposed approaches to the traditional ones as well as comparable results with the state-of-the-art methods.

2014

In this work we were interested in investigating the predictive accuracy of one of the most popular learning schemes for the combination of supervised classification methods: the Stacking Technique proposed by Wolpert (1992) and consolidated by Ting and Witten, (1999) and Seewald (2002). In particular, we made reference to the StackingC (Seewald 2002) as a starting point for our analysis, to which some modifications and extensions were made. Since most of the research on ensembles of classifiers tends to demonstrate that this scheme can perform comparably to the best of the base classifiers as selected by cross-validation, if not better, this motivated us to investigate the performance of the Stacking empirically. An analysis of the results obtained by applying the our Stacking scheme, which includes differences and characteristic implementations compared to what is proposed by the literature, to the set of the dataset generated by means of an experimental design does not lead us to...