• Journal of Intelligence and Information Systems
• /
• v.5 no.1
• /
• pp.95-101
• /
• 1999

In this paper, a comparison between different methods to combine predictions from neural networks will be given. These methods are bagging, bumping, and balancing. Those are based on the analysis of the ensemble generalization error into an ambiguity term and a term incorporating generalization performances of individual networks. Neural Networks and AI machine learning models are prone to overfitting. A strategy to prevent a neural network from overfitting, is to stop training in early stage of the learning process. The complete data set is spilt up into a training set and a validation set. Training is stopped when the error on the validation set starts increasing. The stability of the networks is highly dependent on the division in training and validation set, and also on the random initial weights and the chosen minimization procedure. This causes early stopped networks to be rather unstable: a small change in the data or different initial conditions can produce large changes in the prediction. Therefore, it is advisable to apply the same procedure several times starting from different initial weights. This technique is often referred to as training ensembles of neural networks. In this paper, we presented a comparison of three statistical methods to prevent overfitting of neural network.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.7 no.7
• /
• pp.1705-1720
• /
• 2013

Human face gender recognition requires fast image processing with high accuracy. Existing face gender recognition methods used traditional local features and machine learning methods have shortcomings of low accuracy or slow speed. In this paper, a new framework for face gender recognition to reach fast face gender recognition is proposed, which is based on Local Ternary Pattern (LTP) and Extreme Learning Machine (ELM). LTP is a generalization of Local Binary Pattern (LBP) that is in the presence of monotonic illumination variations on a face image, and has high discriminative power for texture classification. It is also more discriminate and less sensitive to noise in uniform regions. On the other hand, ELM is a new learning algorithm for generalizing single hidden layer feed forward networks without tuning parameters. The main advantages of ELM are the less stringent optimization constraints, faster operations, easy implementation, and usually improved generalization performance. The experimental results on public databases show that, in comparisons with existing algorithms, the proposed method has higher precision and better generalization performance at extremely fast learning speed.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.12
• /
• pp.3063-3068
• /
• 1997

In this paper, we have extended the generalization back-propagation algorithm to multi-layer polynomial higher order neural networks. The purpose of this paper is to describe various pattern recognition using polynomial higher-order neural network. And we have applied shift position T-C test pattern for invariant pattern recognition and measured generalization by mirror symmetry problem. simulation result shows that the ability for invariant pattern recognition increase with the proposed technique. Recognition rate of invariant T-C pattern is 90% effective and of mirror symmetry problem is 70% effective when the proposed technique is utilized. These results are much better than those by the conventional methods.

• International Journal of Advanced Culture Technology
• /
• v.5 no.2
• /
• pp.74-81
• /
• 2017

Recently, there are many companies that use systems based on artificial intelligence. The accuracy of artificial intelligence depends on the amount of learning data and the appropriate algorithm. However, it is not easy to obtain learning data with a large number of entity. Less data set have large generalization errors due to overfitting. In order to minimize this generalization error, this study proposed DGA(Dropout Genetic Algorithm) which can expect relatively high accuracy even though data with a less data set is applied to machine learning based genetic algorithm to deep learning based dropout. The idea of this paper is to determine the active state of the nodes. Using Gradient about loss function, A new fitness function is defined. Proposed Algorithm DGA is supplementing stochastic inconsistency about Dropout. Also DGA solved problem by the complexity of the fitness function and expression range of the model about Genetic Algorithm As a result of experiments using MNIST data proposed algorithm accuracy is 75.3%. Using only Dropout algorithm accuracy is 41.4%. It is shown that DGA is better than using only dropout.

• Journal of the Korean Geographical Society
• /
• v.40 no.4 s.109
• /
• pp.441-453
• /
• 2005

This study aims to generalize the stream network developing algorithm of the geographic hierarchy Stream networks with hierarchy system should be spatially hierarchized in linear features. The generalization procedure of the stream networks are composed of the hierarchy of stream, selection and elimination, and algorithm. Working of stream networks is composed by the decision of direction on stream networks, ranking of stroke segments, and ordering by the strahler method, using geographic data query for controlling selection and elimination of the linear feature by scale. Improved Simoo algorithm was effective in enhancement and decreasing curvature of linear features. Resultantly, it is expected to improve generalization of features with various spatial hierarchy.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.8 no.1
• /
• pp.150-157
• /
• 2004

For the mobile GIS services, it is not acceptable to construct a new map database only for wireless service due to vast cost. But the existing map database cannot be used directly for the services due to narrow wireless bandwidth and shortage of mobile device resources. This thesis proposes spatial data simplification methods, thus, the existing map database enable the mobile GIS services. The proposing methods are based on the existing map generalization techniques. We extend it to mobile environments, and implement. This thesis also includes the issue of discriminative data simplification technique according to display level and map display interface suitable for mobile devices. Research results an estimated by experiments.

• The KIPS Transactions:PartB
• /
• v.12B no.3 s.99
• /
• pp.343-348
• /
• 2005

The goal of this paper is to study the joint effect of factors of neural network teaming procedure. There are many factors, which may affect the generalization ability and teaming speed of neural networks, such as the initial values of weights, the learning rates, and the regularization coefficients. We will apply a constructive training algerian for neural network, then patterns are trained incrementally by considering them one by one. First, we will investigate the effect of these factors on generalization performance and learning speed. Based on these factors' effect, we will propose a joint method that simultaneously considers these three factors, and dynamically hue the learning rate and regularization coefficient. Then we will present the results of some experimental comparison among these kinds of methods in several simulated nonlinear data. Finally, we will draw conclusions and make plan for future work.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.13 no.1
• /
• pp.77-83
• /
• 1995

This paper is concentrated on map generalization in digital environment for automated multi-scale map pro-duction using conventional hardcopy maps. Line generalization is urgently required process to prepare small scale digital map database when large scale map databases are available. This paper outlines a new approach to the line generalization when preparing small scale map on the basis of existing large scale distal map. Line generalizations are conducted based on zero-crossing algorithm using six sheets of 115,000 scale YEOSU area which produced by National Geographic Institute. The results are compared to Douglas-Peucker algorithm and manual method. The study gives full details of the data reduction rates and alternatives based on the proposed algorithm.

• Korean Journal of Acupuncture
• /
• v.17 no.1
• /
• pp.141-156
• /
• 2000

Prognosis in oriental medicine gathers information by four examination methods. It provides important information to understand the degree of deficiency - excessiveness of a patient to treat properly. To generalize the degree of deficiency - excessiveness can be found by seeing the patient's muscle response and pain perception to the palpitations.The theoretical basis to generalize deficiency - excessiveness, oriental and western medical understanding of pain perception and the elasticity of muscle were discussed.The usual symptoms for the excessiveness could include Pain (dislikeness to the palpitation), Stiffness of nape and limbs, Contracture of the limbs, Clonic convulsion and Fast pain. The symptoms for the deficiency could include Pain (likeness to the palpitation, Gastrocnemius muscle spasm, Flaccid paralysis of limbs and Slow pain. More theoretical bases for generalization of deficiency - excessiveness are needed along with the simplifying the complex clinical symptoms. In this way, we can discuss about deficiency - excessiveness with the regard to western medicine to help its generalization.

• Computers and Concrete
• /
• v.11 no.4
• /
• pp.337-350
• /
• 2013

The sensitivity of compressive strength of no-slump concrete to its ingredient materials and proportions, necessitate the use of robust models to guarantee both estimation and generalization features. It was known that the problem of compressive strength prediction owes high degree of complexity and uncertainty due to the variable nature of materials, workmanship quality, etc. Moreover, using the chemical and mineral additives, superimposes the problem's complexity. Traditionally this property of concrete is predicted by conventional linear or nonlinear regression models. In general, these models comprise lower accuracy and in most cases they fail to meet the extrapolation accuracy and generalization requirements. Recently, artificial intelligence-based robust systems have been successfully implemented in this area. In this regard, this paper aims to investigate the use of optimized support vector machine (SVM) to predict the compressive strength of no-slump concrete and compare with optimized neural network (ANN). The results showed that after optimization process, both models are applicable for prediction purposes with similar high-qualities of estimation and generalization norms; however, it was indicated that optimization and modeling with SVM is very rapid than ANN models.