Logarithmic learning for generalized classifier neural network

Özyildirim, Buse Melis; Avcı, Mutlu

doi:10.1016/j.neunet.2014.08.004

Cited by 9 publications

(5 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The value of critical sensitivity is above ninety percent in both cases. The reached maximum accuracy value is fully comparable to other presented results (Kulluk et al 2012;Ozyildirim and Avci 2014;Asafuddoula et al 2017;Yin and Gelenbe 2018) summarized in Tables 4 and 5 along with reached minimum and maximal accuracy together with placement among benchmark techniques.…”

Section: Case Study: Iris Flower Classification Tasksupporting

confidence: 87%

“…As seen in Tables 4 and 5 our method is comparable with standard techniques of classification. To see how competitive our results are, we went through several papers on this topic (Kulluk et al 2012;Ozyildirim and Avci 2014;Rani and Ganesh 2014;Abdar et al 2017;Asafuddoula et al 2017;Kahramanli 2017;Aslan et al 2018;Li and Chen 2018;Talabni and Engin 2018;Yin and Gelenbe 2018;Austria et al 2019;Chan and Chin 2019;Kraipeerapun and Amornsamankul 2019;Rahman et al 2020). The rank of our novel method has been evaluated for every dataset and during both training and cross-validation processes.…”

Section: Application To Other Pattern Setsmentioning

confidence: 99%

“…For training, we compared the results with approximator using a spiking random neural network using five other benchmarks (Yin and Gelenbe 2018). For cross-validation results comparison we used generalized classifier neural network and its logarithmic learning implementation, probabilistic neural network and standard multilayer perceptron as presented in Ozyildirim and Avci (2014). The results of an incremental ensemble classifier method (Asafuddoula et al 2017) were also used as a benchmark.…”

Section: Application To Other Pattern Setsmentioning

confidence: 99%

See 2 more Smart Citations

Concept of hidden classes in pattern classification

Hrebik

Kukal

2023

Artif Intell Rev

View full text Add to dashboard Cite

Our paper presents a novel approach to pattern classification. The general disadvantage of a traditional classifier is in too different behaviour and optimal parameter settings during training on a given pattern set and the following cross-validation. We describe the term critical sensitivity, which means the lowest reached sensitivity for an individual class. This approach ensures a uniform classification quality for individual class classification. Therefore, it prevents outlier classes with terrible results. We focus on the evaluation of critical sensitivity, as a quality criterion. Our proposed classifier eliminates this disadvantage in many cases. Our aim is to present that easily formed hidden classes can significantly contribute to improving the quality of a classifier. Therefore, we decided to propose classifier will have a relatively simple structure. The proposed classifier structure consists of three layers. The first is linear, used for dimensionality reduction. The second layer serves for clustering and forms hidden classes. The third one is the output layer for optimal cluster unioning. For verification of the proposed system results, we use standard datasets. Cross-validation performed on standard datasets showed that our critical sensitivity-based classifier provides comparable sensitivity to reference classifiers.

show abstract

Section: Case Study: Iris Flower Classification Tasksupporting

confidence: 87%

Section: Application To Other Pattern Setsmentioning

confidence: 99%

Section: Application To Other Pattern Setsmentioning

confidence: 99%

See 1 more Smart Citation

Concept of hidden classes in pattern classification

Hrebik

Kukal

2023

Artif Intell Rev

View full text Add to dashboard Cite

show abstract

“…To manage those problems, the latest study offers a solution using a Logarithmic learning generalized classifier neural network (LGCNN) [20], [21], [27]. The difference of this method with other RBF neural network is in using a logarithmic function to calculate the error for each epoch [12]. With this function, it will decrease the iteration to get the minimum error.…”

Section: Generalized Classifier Neural Networkmentioning

confidence: 99%

Improvement on I-Devices Using L-GCNN Classifier for Smart Mosque Simulation

Fadila¹,

Hariyadi²,

Hanani³

et al. 2022

ijasca

View full text Add to dashboard Cite

I-Device (Intelligent Devices) is one of the fastest growing devices since the beginning of this decade. Some of its major problems are accuracy and performance. This study aims to present an improvement in the performance of those devices. We used a simulation application for I-Devices to conduct the experiment. The simulation was built based on classifying results using Logarithmic learning for Generalized Classifier Neural Networks (L-GCNN). The output was a simulation that will be implemented on a smart mosque system. L-GCNN itself was a modification method of GCNN to improve the processing speed and have high accuracy as a classifier method. This method will take a role when the given parameters meet the conditions of the devices to take an action. To simplify the understanding of the simulation models, we used a game application to make an interactive simulation for our project in an environment that represents the real-world condition of the mosque. The result of this study shows that the devices could make a decision by themselves accurately. Additionally, using LGCNN models, we could reduce the processing iteration compared to other models. The experiment results show that LGCNN has an average value of 90% in accuracy, precision, recall, and f1. Keywords: Automation, Classifier, L-GCNN, Neural Network, Decision.

show abstract

“…In recent years, a large number of experts have done a lot of works on this issue. Such as in [14], a logarithmic learning method for generalized classifier neural network was presented to handle the convergence problem and improve the rate of convergence when the initial smoothing parameter value deviated from the optimal one. Moreover, an adaptive state-feedback controller based on neural networks for a class of high-order stochastic uncertain systems was designed to eliminate the unknown nonlinearities restrictions in [15].…”

Section: Introductionmentioning

confidence: 99%

A novel adaptive neural network constrained control for solid oxide fuel cells via dynamic anti-windup

Liu

2016

Neurocomputing

View full text Add to dashboard Cite

Logarithmic learning for generalized classifier neural network

Cited by 9 publications

References 11 publications

Concept of hidden classes in pattern classification

Concept of hidden classes in pattern classification

Improvement on I-Devices Using L-GCNN Classifier for Smart Mosque Simulation

A novel adaptive neural network constrained control for solid oxide fuel cells via dynamic anti-windup

Contact Info

Product

Resources

About