• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.31B no.4
• /
• pp.145-154
• /
• 1994

This paper proposes an efficient method for improving the training performance of the neural network by using a hybrid of a stochastic approximation and a backpropagation algorithm. The proposed method improves the performance of the training by appliying a global optimization method which is a hybrid of a stochastic approximation and a backpropagation algorithm. The approximate initial point for a stochastic approximation and a backpropagation algorihtm. The approximate initial point for fast global optimization is estimated first by applying the stochastic approximation, and then the backpropagation algorithm, which is the fast gradient descent method, is applied for a high speed global optimization. And further speed-up of training is made possible by adjusting the training parameters of each of the output and the hidden layer adaptively to the standard deviation of the neuron output of each layer. The proposed method has been applied to the parity checking and the pattern classification, and the simulation results show that the performance of the proposed method is superior to that of the backpropagation, the Baba's MROM, and the Sun's method with randomized initial point settings. The results of adaptive adjusting of the training parameters show that the proposed method further improves the convergence speed about 20% in training.

• The KIPS Transactions:PartB
• /
• v.14B no.3 s.113
• /
• pp.153-162
• /
• 2007

Sasang medicine is peculiar medicine that constitution of a human classify four types and differ treatment method by physical constitution. In this way the most important thing is very difficult problem that classification of Sasang constitution and discriminate correctly. Therefore, in this paper targets diagnosis medical appliances development of hybrid form that can behave constitution classification and sees among for this paper to propose about method to grasp characteristic that is morphology about eye, nose, ear and mouth be based on appearance and manner of speaking. In this paper, classified and verified this for Sasang constitution through the QSCC II program at 1 step and present method that more exactly and conveniently analyzing measure each physical constitution feature by survey about eye, nose, ear and mouth at 2 steps. Also, extraction and analyze and verified main area of physical constitution classification based on front face and side face at 3 steps. Such propose method to extraction the principal face region based on face color from front face and side face for correct physical constitution classification diagnosis appliance development through experiment consideration and verification process.

• Journal of Digital Contents Society
• /
• v.17 no.6
• /
• pp.565-579
• /
• 2016

In this paper a novel cancer subtype's classifier based on Hybrid Samples Balanced Genetic Algorithm with Extreme Learning Machine (hSBGA-ELM) is presented. Proposed cancer subtype's classifier uses genes' expression data of 16063 genes from open Global Cancer Map (GCM) data base for accurate cancer subtype's classification. Proposed method efficiently classifies 14 subtypes of cancer (breast, prostate, lung, colorectal, lymphoma, bladder, melanoma, uterus, leukemia, renal, pancreas, ovary, mesothelioma and CNS). Proposed hSBGA-ELM unifies genes' selection procedure and cancer subtype's classification into one framework. Proposed Hybrid Samples Balanced Genetic Algorithm searches a reduced robust set of genes responsible for cancer subtype's classification from 16063 genes available in GCM data base. Selected reduced set of genes is used to build cancer subtype's classifier using Extreme Learning Machine (ELM). As a result, reduced set of robust genes guarantees stable generalization performance of the proposed cancer subtype's classifier. Proposed hSBGA-ELM discovers 95 genes probably responsible for cancer. Comparison with existing cancer subtype's classifiers clear indicates efficiency of the proposed method.

• Journal of the Korean Data and Information Science Society
• /
• v.12 no.2
• /
• pp.11-25
• /
• 2001

There are several algorithms for classification in modeling relations, patterns, and rules which exist in data. We learn to classify objects on the basis of instances presented to us, not by being given a set of classification rules. The Bayesian learning uses the probability distribution to express our knowledge about unknown parameters and update our knowledge by the law of probability as the evidence gathered from data. Also, the neural network models are designed for predicting an unknown category or quantity on the basis of known attributes by training. In this paper, we compare the misclassification error rates of Bayesian Neural Network method with those of other classification algorithms, CHAID, CART, and QUBST using several data sets.

• International Journal of Computer Science & Network Security
• /
• v.22 no.4
• /
• pp.193-202
• /
• 2022

The classification of hyperspectral imagery (HSI) is essential in the surface of earth observation. Due to the continuous large number of bands, HSI data provide rich information about the object of study; however, it suffers from the curse of dimensionality. Dimensionality reduction is an essential aspect of Machine learning classification. The algorithms based on feature extraction can overcome the data dimensionality issue, thereby allowing the classifiers to utilize comprehensive models to reduce computational costs. This paper assesses and compares two HSI classification techniques. The first is based on the Joint Spatial-Spectral Stacked Autoencoder (JSSSA) method, the second is based on a shallow Artificial Neural Network (SNN), and the third is used the SVM model. The performance of the JSSSA technique is better than the SNN classification technique based on the overall accuracy and Kappa coefficient values. We observed that the JSSSA based method surpasses the SNN technique with an overall accuracy of 96.13% and Kappa coefficient value of 0.95. SNN also achieved a good accuracy of 92.40% and a Kappa coefficient value of 0.90, and SVM achieved an accuracy of 82.87%. The current study suggests that both JSSSA and SNN based techniques prove to be efficient methods for hyperspectral classification of snow features. This work classified the labeled/ground-truth datasets of snow in multiple classes. The labeled/ground-truth data can be valuable for applying deep neural networks such as CNN, hybrid CNN, RNN for glaciology, and snow-related hazard applications.

• Korean Journal of Fisheries and Aquatic Sciences
• /
• v.56 no.5
• /
• pp.708-715
• /
• 2023

Hybridization is a major production method used to combine beneficial traits from two different species to obtain a potentially dominant trait. In China, Takifugu obscurus and T. rubripes were artificially crossed, and the resulting hybrids had an average body weight 38.06-8.93% higher than that of the parental species, which enabled the hybrids to be grown in freshwater. This study aimed to provide the basic data necessary for the classification of T. obscurus♀×T. rubripes♂ hybrids in terms of economic value and market potential. Morphological comparing the morphology of hybrids and parental species, we discovered that the hybrids had intermediate traits of the parental species. In morphometrics, the hybrid index (HI) value of head length against standard length was close to the trait of T. rubripes, and the HI values of preanal length and predorsal length were close to those of T. obscurus; however, the HI values of nasal length, snout length, length of anal fin, length of pectoral fin, caudal peduncle depth and caudal peduncle length were found to be unique characteristics of the hybrids. Regarding molecular analysis, a 99.8% nucleotide sequence similarity was found between the hybrid and T. obscurus.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.39 no.1
• /
• pp.105-115
• /
• 2016

Ensemble classification involves combining individually trained classifiers to yield more accurate prediction, compared with individual models. Ensemble techniques are very useful for improving the generalization ability of classifiers. The random subspace ensemble technique is a simple but effective method for constructing ensemble classifiers; it involves randomly drawing some of the features from each classifier in the ensemble. The instance selection technique involves selecting critical instances while deleting and removing irrelevant and noisy instances from the original dataset. The instance selection and random subspace methods are both well known in the field of data mining and have proven to be very effective in many applications. However, few studies have focused on integrating the instance selection and random subspace methods. Therefore, this study proposed a new hybrid ensemble model that integrates instance selection and random subspace techniques using genetic algorithms (GAs) to improve the performance of a random subspace ensemble model. GAs are used to select optimal (or near optimal) instances, which are used as input data for the random subspace ensemble model. The proposed model was applied to both Kaggle credit data and corporate credit data, and the results were compared with those of other models to investigate performance in terms of classification accuracy, levels of diversity, and average classification rates of base classifiers in the ensemble. The experimental results demonstrated that the proposed model outperformed other models including the single model, the instance selection model, and the original random subspace ensemble model.

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.1106-1110
• /
• 2004

A human sign language is generally composed of both static and dynamic gestures. Each gesture is represented by a hand shape, its position, and hand movement (for a dynamic gesture). One of the problems found in automated sign language translation is on segmenting a hand movement that is part of a transitional movement from one hand gesture to another. This transitional gesture conveys no meaning, but serves as a connecting period between two consecutive gestures. Based on the observation that many dynamic gestures as appeared in Thai sign language dictionary are of quasi-periodic nature, a method was developed to differentiate between a (meaningful) dynamic gesture and a transitional movement. However, there are some meaningful dynamic gestures that are of non-periodic nature. Those gestures cannot be distinguished from a transitional movement by using the signal quasi-periodicity. This paper proposes a hybrid method using a combination of the periodicity-based gesture segmentation method with a HMM-based gesture classifier. The HMM classifier is used here to detect dynamic signs of non-periodic nature. Combined with the periodic-based gesture segmentation method, this hybrid scheme can be used to identify segments of a transitional movement. In addition, due to the use of quasi-periodic nature of many dynamic sign gestures, dimensionality of the HMM part of the proposed method is significantly reduced, resulting in computational saving as compared with a standard HMM-based method. Through experiment with real measurement, the proposed method's recognition performance is reported.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.21 no.11
• /
• pp.997-1004
• /
• 2011

A fault detection algorithm of a charge and discharge system to ensure the safe use of hybrid electric vehicle is proposed in this paper. This algorithm can be used as a complementary way to existing fault detection technique for a charge and discharge system. The proposed algorithm uses a SVDD technique, which additionally utilizes two methods for learning a large amount of data; one is to incrementally learn a large amount of data, the other one is to remove the data that does not affect the next learning using a new data reduction technique. Removal of data is selected by using lines connecting support vectors. In the proposed method, the data processing speed is drastically improved and the storage space used is remarkably reduced than the conventional methods using the SVDD technique only. A battery data and speed data of a commercial hybrid electrical vehicle are utilized in this study. A fault boundary is produced via SVDD techniques using the input and output in normal operation of the system without using mathematical modeling. A fault detection simulation is performed using both an artificial fault data and the obtained fault boundary via SVDD techniques. In the fault detection simulation, fault detection time via proposed algorithm is compared with that of the peak-peak method. Also the proposed algorithm is revealed to detect fault in the region where conventional peak-peak method is never able to do.

• Journal of Advanced Marine Engineering and Technology
• /
• v.40 no.2
• /
• pp.138-145
• /
• 2016

This paper proposes an improved tabu search method for subset selection in multiple linear regression models. Variable selection is a vital combinatorial optimization problem in multivariate statistics. The selection of the optimal subset of variables is necessary in order to reliably construct a multiple linear regression model. Its applications widely range from machine learning, timeseries prediction, and multi-class classification to noise detection. Since this problem has NP-complete nature, it becomes more difficult to find the optimal solution as the number of variables increases. Two typical metaheuristic methods have been developed to tackle the problem: the tabu search algorithm and hybrid genetic and simulated annealing algorithm. However, these two methods have shortcomings. The tabu search method requires a large amount of computing time, and the hybrid algorithm produces a less accurate solution. To overcome the shortcomings of these methods, we propose an improved tabu search algorithm to reduce moves of the neighborhood and to adopt an effective move search strategy. To evaluate the performance of the proposed method, comparative studies are performed on small literature data sets and on large simulation data sets. Computational results show that the proposed method outperforms two metaheuristic methods in terms of the computing time and solution quality.