• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.363-363
• /
• 2023

Artificial Neural Networks (ANN) is a powerful tool for addressing various practical problems and it has been extensively applied in areas of water resources. In this study, Artificial Neural Networks (ANNs) were developed for flood forecasting at specific locations on the Han River. The Partial Mutual Information (PMI) technique was used to select input variables for ANNs that are neither over-specified nor under-specified while adequately describing the underlying input-output relationships. Historical observations including discharges at the Paldang Dam, flows from tributaries, water levels at the Paldang Bridge, Banpo Bridge, Hangang Bridge, and Junryu gauge station, and time derivatives of the observed water levels were considered as input candidates. Lagged variables from current time t to the previous five hours were assumed to be sufficient in this study. A three-layer neural network with one hidden layer was used and the neural network was optimized by selecting the optimal number of hidden neurons given the selected inputs. Given an ANN architecture, the weights and biases of the network were determined in the model training. The use of PMI-based input variable selection and optimized ANNs for different sites were proven to successfully predict water levels during flood periods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.3
• /
• pp.528-550
• /
• 2024

Warehousing demand prediction is an essential part of the supply chain, providing a fundamental basis for product manufacturing, replenishment, warehouse planning, etc. Existing forecasting methods cannot produce accurate forecasts since warehouse demand is affected by external factors such as holidays and seasons. Some aspects, such as consumer psychology and producer reputation, are challenging to quantify. The data can fluctuate widely or do not show obvious trend cycles. We introduce a new model for warehouse demand prediction called MAGRU, which stands for Multi-layer Attention with GRU. In the model, firstly, we perform the embedding operation on the input sequence to quantify the external influences; after that, we implement an encoder using GRU and the attention mechanism. The hidden state of GRU captures essential time series. In the decoder, we use attention again to select the key hidden states among all-time slices as the data to be fed into the GRU network. Experimental results show that this model has higher accuracy than RNN, LSTM, GRU, Prophet, XGboost, and DARNN. Using mean absolute error (MAE) and symmetric mean absolute percentage error(SMAPE) to evaluate the experimental results, MAGRU's MAE, RMSE, and SMAPE decreased by 7.65%, 10.03%, and 8.87% over GRU-LSTM, the current best model for solving this type of problem.

• Transactions of Materials Processing
• /
• v.33 no.2
• /
• pp.96-102
• /
• 2024

This study addresses evaluation of performance of hardening model for a titanium alloy (Ti6Al4V) based on the artificial neural network (ANN) regarding the strain rate and the temperature. Uniaxial compression tests were carried out at different strain rates from 0.001 /s to 10 /s and temperatures from 575 ℃ To 975 ℃. Using the experimental data, ANN models were trained and tested with different hyperparameters, such as size of hidden layer and optimizer. The input features were determined with the equivalent plastic strain, strain rate, and temperature while the output value was set to the equivalent stress. When the number of data is sufficient with a smooth tendency, both the Bayesian regulation (BR) and the Levenberg-Marquardt (LM) show good performance to predict the flow behavior. However, only BR algorithm shows a predictability when the number of data is insufficient. Furthermore, a proper size of the hidden layer must be confirmed to describe the behavior with the limited number of the data.

• Journal of the Korea Society of Computer and Information
• /
• v.5 no.2
• /
• pp.30-36
• /
• 2000

The hidden node plays a role of the functional units that classifies the features of input pattern in the given question. Therefore, a neural network that consists of the number of a suitable optimum hidden node has be on the rise as a factor that has an important effect upon a result. However there is a problem that decides the number of hidden nodes based on back-propagation learning algorithm. If the number of hidden nodes is designated very small perfect learning is not done because the input pattern given cannot be classified enough. On the other hand, if designated a lot, overfitting occurs due to the unnecessary execution of operation and extravagance of memory point. So, the recognition rate is been law and the generality is fallen. Therefore, this paper suggests a method that decides the number of neural network node with feature information consisted of the parameter of learning algorithm. It excludes a node in the Pruning target, that has a maximum value among the feature value obtained and compares the average of the rest of hidden node feature value with the feature value of each hidden node, and then would like to improve the learning speed of neural network deciding the optimum structure of the multi-layer neural network as pruning the hidden node that has the feature value smaller than the average.

• Journal of Mechanical Science and Technology
• /
• v.20 no.12
• /
• pp.2079-2086
• /
• 2006

A fundamental study for developing a fault diagnosis system of a pump is performed by using neural network. Acoustic signals were obtained and converted to frequency domain for normal products and artificially deformed products. The neural network model used in this study was 3-layer type composed of input, hidden, and output layer. The normalized amplitudes at the multiples of real driving frequency were chosen as units of input layer. And the codes of pump malfunctions were selected as units of output layer. Various sets of teach signals made from original data by eliminating some random cases were used in the training. The average errors were approximately proportional to the number of untaught data. Neural network trained by acoustic signals can detect malfunction or diagnose fault of a given machine from the results.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.19 no.11
• /
• pp.2264-2275
• /
• 1994

In this paper, image compression based on neural networks is presented with block classification and coding. Multilayer neural networks with error back-propagation learning algorithm are used to transform the normalized image date into the compressed hidden values by reducing spatial redundancies. Image compression can basically be achieved with smaller number of hidden neurons than the numbers of input and output neurons. Additionally, the image blocks can be grouped for adaptive compression rates depending on the characteristics of the complexity of the blocks in accordance with the sensitivity of the human visual system(HVS). The quantized output of the hidden neuron can also be entropy coded for an efficient transmission. In computer simulation, this approach lie in the good performances even with images outside the training set and about 25:1 compression rate was achieved using the entropy coding without much degradation of the reconstructed images.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09a
• /
• pp.599-602
• /
• 2003

A new technique is presented to construct predictive models of plasma etch processes. This was accomplished by combining a backpropagation neural network (BPNN) and a random generator (RC). The RG played a critical role to control neuron gradients in the hidden layer, The predictive model constructed in this way is referred to as a randomized BPNN (RG-BPNN). The proposed scheme was evaluated with a set of experimental plasma etch process data. The etch process was characterized by a 2$^3$ full factorial experiment. The etch responses modeled are 4, including aluminum (Al) etch rate, profile angle, Al selectivity, and do bias. Additional test data were prepared to evaluate model appropriateness. The performance of RC-BPNN was evaluated as a function of the number of hidden neurons and the range of gradient. for given range and hidden neurons, 100 sets of random neuron gradients were generated and among them one best set was selected for evaluation. Compared to the conventional BPNN, the proposed RC-BPNN demonstrated about 50％ improvements in all comparisons. This illustrates that the RG-BPNN of multi-valued gradients is an effective way to considerably improve the predictive ability of current BPNN of single-valued gradient.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.25 no.1
• /
• pp.97-104
• /
• 2015

An 'associative cube', a class of auto-associative memories, is revisited here, in which training data and hidden orthogonal basis functions such as wavelet packets or Fourier kernels, are combined in the weight cube. This weight cube has hidden units in its depth, represented by a three dimensional cubic structure. We develop an unsupervised incremental learning mechanism based upon the adaptive least squares method. Training data are mapped into orthogonal basis vectors in a least-squares sense by updating the weights which minimize an energy function. Therefore, a prescribed orthogonal kernel is incrementally assigned to an incoming data. Next, we show how a decoding procedure finds the closest one with a competitive network in the hidden layer. As noisy test data are applied to an associative cube, the nearest one among the original training data are restored in an optimal sense. The simulation results confirm robustness of associative cubes even if test data are heavily distorted by various types of noise.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.5 no.4
• /
• pp.684-702
• /
• 2011

This paper proposes a TDMA-based multichannel medium access control (MAC) protocol for QoS provisioning in mobile ad hoc networks (MANETs) that enables nodes to transmit their packets in distributed channels. The IEEE 802.11 standard supports multichannel operation at the physical (PHY) layer but its MAC protocol is designed only for a single channel. The single channel MAC protocol does not work well in multichannel environment because of the multichannel hidden terminal problem. Our proposed protocol enables nodes to utilize multiple channels by switching channels dynamically, thus increasing network throughput. Although each node of this protocol is equipped with only a single transceiver but it solves the multichannel hidden terminal problem using temporal synchronization. The proposed energy efficient multichannel MAC (EM-MAC) protocol takes the advantage of both multiple channels and TDMA, and achieves aggressive power savings by allowing nodes that are not involved in communications to go into power saving "sleep mode". We consider the problem of providing QoS guarantee to nodes as well as to maintain the most efficient use of scarce bandwidth resources. Our scheme improves network throughput and lifetime significantly, especially when the network is highly congested. The simulation results show that our proposed scheme successfully exploits multiple channels and significantly improves network performance by providing QoS guarantee in MANETs.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.7
• /
• pp.1773-1780
• /
• 1996

KBANN (knowledge-based artificial neural network) combining the symbolic approach and the numerical approach has been shown to be more effective than other machine learning models. However KBANN doesn't have the theory refinement ability because the topology of network can't be altered dynamically. Although TopGen was proposed to extend the ability of KABNN in this respect, it also had some defects due to the link-ing of hidden nodes to input nodes and the use of beam search. The algorithm which could solve this TopGen's defects, by adding the hidden nodes linked to next layer nodes and using hill-climbing search with backtracking, is designed.