This paper firstly introduces the basic concept of covering algorithm and kernel covering algorithm (KCA) adopting kernel function, then analyzes the influence of proximity principle used to judge rejection points on ...
详细信息
This paper firstly introduces the basic concept of covering algorithm and kernel covering algorithm (KCA) adopting kernel function, then analyzes the influence of proximity principle used to judge rejection points on classifier’s effect. FKCA, i.e. Fuzzy Kernel covering algorithm, is proposed to improve the performance of classifier. The main improvement of FKCA is the change of radius selection and introduction of membership function. Also we discuss the influence of isolated covering and provide a method to reduce the number of coverings by the combination of second scanning with contribution calculation. Experiments show the performance gap between KCA and FKCA, and comparisons with other fuzzy classifiers are also performed. We apply FKCA to character recognition of car plates and the result is satisfactory.
This paper firstly introduces the basic concept of covering algorithm and kernel covering algorithm(KCA) adopting kernel function,then analyzes the influence of proximity principle used to judge rejection points on cl...
详细信息
This paper firstly introduces the basic concept of covering algorithm and kernel covering algorithm(KCA) adopting kernel function,then analyzes the influence of proximity principle used to judge rejection points on classifier's ***,*** Kernel covering algorithm,is proposed to improve the performance of *** main improvement of FKCA is the change of radius selection and introduction of membership *** we discuss the influence of isolated covering and provide a method to reduce the number of coverings by the combination of second scanning with contribution *** show the performance gap between KCA and FKCA,and comparisons with other fuzzy classifiers are also *** apply FKCA to character recognition of car plates and the result is satisfactory.
The covering algorithm is an often-used rule induction method. The main shortcoming of this algorithm is its inability to handle continuous-type data. The present research proposes a novel method that integrates genet...
详细信息
The covering algorithm is an often-used rule induction method. The main shortcoming of this algorithm is its inability to handle continuous-type data. The present research proposes a novel method that integrates genetic algorithms with covering algorithms in support of rule induction dealing with both continuous and categorical data types. We illustrate this method and demonstrate its effectiveness with data obtained directly from the flight data recorders of Boeing 747-400 airplanes. The results indicate that the adapted covering algorithm is feasible as a complete rule induction method. (C) 2003 Elsevier Ltd. All rights reserved.
covering algorithms are very useful learning methods of neural networks, and their computing complexity is lower than that of the learning method based on search mechanism or programming-based learning algorithm. Cove...
详细信息
covering algorithms are very useful learning methods of neural networks, and their computing complexity is lower than that of the learning method based on search mechanism or programming-based learning algorithm. covering algorithms not only are applied to deal with vast data set but also provide a new constructive learning method of neural networks. However, they are based on the assumption that all of the training samples are accurate and the instance that some of the training samples are not accurate is not discussed. If the methods are applied to no-accurate data directly, the result is not satisfactory. The research discussed the problem of forecasting time series where there are some no-accurate data. The covering algorithm improved and the definitions of covering intensity and no-acknowledge sample are introduced. The improved covering algorithm is called a structural learning algorithm (SLA). SLA is applied to forecasting a time series which is composed of Shanghai's stock integrating index, and the satisfying results are achieved. It is expected that SLA will have wide applications.
This paper puts forward to using granularity calculating of quotient space theory dealing with classification problems about machine learning. According to the prior knowledge or the clustering of the training example...
详细信息
ISBN:
(纸本)0780378652
This paper puts forward to using granularity calculating of quotient space theory dealing with classification problems about machine learning. According to the prior knowledge or the clustering of the training examples, the training data reorganized with granularity to form new instances, and to learn from the new instances. The different sorts of instances, which are combined in granularity, will be classified through different layers of classifiers. By this way, the difficulty of the learning is reduced, and the capacity of the learning from the instances is increased, and the classifying accuracy is improved. At the same time, the method can recognize the different sorts of instances, which features are very similar, and improve its generalization of recognition, and reduce the complicacy of calculating. The detailed procedures of the method using covering algorithm and its experimental results are presented. The results show that the method is effective.
This paper presents RULES-5, a new induction algorithm for effectively handling problems involving continuous attributes. RULES-5 is a 'covering' algorithm that extracts IF-THEN rules from examples presented t...
详细信息
This paper presents RULES-5, a new induction algorithm for effectively handling problems involving continuous attributes. RULES-5 is a 'covering' algorithm that extracts IF-THEN rules from examples presented to it. The paper first reviews existing methods of rule extraction and dealing with continuous attributes. It then describes the techniques adopted for RULES-5 and gives a step-by-step example to illustrate their operation. The paper finally gives the results of applying RULES-5 and other algorithms to benchmark problems. These clearly show that RULES-5 generates rule sets that are more accurate than those produced by its immediate predecessor RULES-3 Plus and by a well-known commercially available divide-and-conquer machine learning algorithm.
In this paper, a new covering algorithm called FCV1 is presented. FCV1 comprises two algorithms, one of which is able to fast search for a partial rule and exclude the larg portion of negative examples, the other algo...
详细信息
In this paper, a new covering algorithm called FCV1 is presented. FCV1 comprises two algorithms, one of which is able to fast search for a partial rule and exclude the larg portion of negative examples, the other algorithm incorporates the more optimized greedy set-covering algorithm, and runs on a small portion of training examples. Hence,the training process of FCV1 is much faster than that of AQ15.
The empirical inductive algorithms that utilize the covering paradigm (such as the AQx and CNx families of inductive systems) comprise various heuristics and statistical tools so that the core of the covering paradigm...
详细信息
The empirical inductive algorithms that utilize the covering paradigm (such as the AQx and CNx families of inductive systems) comprise various heuristics and statistical tools so that the core of the covering paradigm remains often quite hidden. The goal of this paper is thus to disclose theoretical underlying principles of covering learning algorithms. By exploiting the set theory, the paper exhibits how the correctness and generality required for decision rules induced by a covering algorithm may be satisfied. The principle differences between a genuine theoretical approach and actual empirical machine learning algorithms are also discussed.
作者：
FRIEDLI, AETH
SEMINAR ANGEW MATHZURICH 8006SWITZERLAND
covering algorithms utilized in methods of search for finding zeros of complex polynomials are investigated from both a deterministic and a probabilistic point of view. q-coverings are found to be optimal, and numeric...
详细信息
covering algorithms utilized in methods of search for finding zeros of complex polynomials are investigated from both a deterministic and a probabilistic point of view. q-coverings are found to be optimal, and numerical examples of this and other types of coverings are given.
暂无评论