• Journal of Korea Multimedia Society
• /
• v.6 no.7
• /
• pp.1312-1320
• /
• 2003

In this paper, we consider a multilayer neural network, with a single hidden layer. Error backpropagation learning method used widely in multilayer neural networks has a possibility of local minima due to the inadequate weights and the insufficient number of hidden nodes. So we propose a fuzzy supervised learning algorithm by using self-generation that self-generates hidden nodes by the compound fuzzy single layer perceptron and modified ART1. From the input layer to hidden layer, a modified ART1 is used to produce nodes. And winner take-all method is adopted to the connection weight adaptation, so that a stored pattern for some pattern gets updated. The proposed method has applied to the student identification card images. In simulation results, the proposed method reduces a possibility of local minima and improves learning speed and paralysis than the conventional error backpropagation learning algorithm.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.2C
• /
• pp.31-40
• /
• 2005

we propose an enhanced self-generation supervised algorithm that by combining an ART algorithm and the delta-bar-delta method. Form the input layer to the hidden layer, ART-1 and ART-2 are used to produce nodes, respectively. A winner-take-all method is adopted to the connection weight adaption so that a stored pattern for some pattern is updated. we test the recognition of student identification, a certificate of residence, and an identifier from container that require nodes of hidden layers in neural network. In simulation results, the proposed self-generation supervised learning algorithm reduces the possibility of local minima and improves learning speed and paralysis than conventional neural networks.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.9 no.4
• /
• pp.883-889
• /
• 2005

Error backpropagation algorithm of multilayer perceptron may result in local-minima because of the insufficient nodes in the hidden layer, inadequate momentum set-up, and initial weights. In this paper, we proposed the ART-1 based fuzzy supervised learning algorithm which is composed of ART-1 and fuzzy single layer supervised learning algorithm. The Proposed fuzzy supervised learning algorithm using self-generation method applied not only ART-1 to creation of nodes from the input layer to the hidden layer, but also the winer-take-all method, modifying stored patterns according to specific patterns. to adjustment of weights. We have applied the proposed learning method to the problem of recognizing a resident registration number in resident cards. Our experimental result showed that the possibility of local-minima was decreased and the teaming speed and the paralysis were improved more than the conventional error backpropagation algorithm.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.43 no.5
• /
• pp.861-865
• /
• 1994

By using Neural Networks, a 8$\times$8 Banyan network controller is designed and implemented. In order to solve internal blocking and output blocking, Winner-Take-All method is used. The longer queue takes higher priority. First-in-first-out method is used among the non-blocking cells in the queue selected.The required time to select a cell is 2.7 $\mu$sec for 155Mbps. The implemented controller using Xilinx FPGA chip selects cells within 2.5$\mu$sec.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.32B no.10
• /
• pp.1347-1357
• /
• 1995

In this paper, we describe a modularized neuroclassifier for enhancing the recognition accuracy of mixed printed and handwritten numerals. This classifier combines four modularized subclassifiers using multi-layer perceptron module. The input of each subclassifier is comprised of a group of specialized feature sets. On applying this method to combining several subclassifiers for unconstrained handwritten numerals, the experimental result shows that the performance of individual subclassifier can be improved. In winner-take-all voting method, the result of subclassifier having the highest RF value is selected as the output. The generality of this classifier is tested with 1,080 printed and 3,000 handwritten numerals that was not shown in training the neural networks. Experimental results show 98.2% recognition rate. The typical recognition test with a threshold value(RF=1.5) has shown 97% recognition, 1% substitution and 2% rejection rates.

• The Journal of Fisheries Business Administration
• /
• v.29 no.2
• /
• pp.111-141
• /
• 1998

Why are auctions so prevalent in fisheries fields\ulcorner One answer is, perhaps, that fisheries products have no standard value. The price of any catch of fish (at least of fish destined for the fresh fish market) depends on the demand and supply conditions at a specific moment of time, influenced by prospective market developments and prices must be remade for each transaction. There are various auction schemes including written-bid method in Korea. It is difficult to make decision the application of auction selling in market distribution. One reason is the absence of adequate data on which to base firm statements. There is very little precise information about the relative volume handled by auction scheme as compared with other sales schemes Because of the paucity reliable data, one must depend largely on qualitative in attempting to above this problem. This paper is designed to examine which auction scheme is more efficient as a economic tool by introducing the three auctions, as samples, utilized in Koheung peninsula. Koheung coast with abundant fisheries resource, locates on the Middle-South part in Korea and has three auction firm operating by fisheries cooperatives. fur selling of dry fish included sea weed, live fish and fresh fish respectively, As a result, 1 found that there are three interesting auction schemes as follow : 1. More than one winners are selected as buyers. The highest bidder can at first get fishes he wants to buy, second winner can get surplus fishes after the highest bidder employes one's privilege with the highest pay, and next winner would be a buyer if fishes would be left in sequence after being sold with higher price. Every fisherman can sell one' s fishes with equal unit price if he delivery it to the winner within one day. Therefore, all the vendors feel they are equal members of fisheries cooperation. 2. Written-bid pricing on the cover of handy book. It is easy to write and to erase the figure more than on the small black board, and is convenient also to keep in the pocket. 3. Auctioning on the fixed platform with fixed fish tank is a very fast auction scheme in spite of short displaying time. Auctioneer presides bidding at one place on the fixed platform, instead of moving, vendors should carry a container of live fishes in the fish tank into showing table in front of would-be buyers and auctioneer. Although the applicability of the auction system to a marketing problem depends in part on subjective considerations by those making decision, basically it is a matter of comparative economic efficiency. In general, if the scheme maximizes returns in relation to the effort expanded by both buyers and sellers, it will be utilized. If it does not, a more efficiency may take place over a period of time, but, even more important, those making decisions may become aware of the potentialities of new schemes. Therefore, in order to applicate the three interesting auction schemes introduced in this paper to other fisheries market, it is necessary not only to analyze many other auction schemes but also to compare the economic efficiency those schemes utilizing in other fisheries market.