• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.4
• /
• pp.1275-1292
• /
• 2021

As a fine-grained classification problem, aspect-level sentiment classification predicts the sentiment polarity for different aspects in context. To address this issue, researchers have widely used attention mechanisms to abstract the relationship between context and aspects. Still, it is difficult to effectively obtain a more profound semantic representation, and the strong correlation between local context features and the aspect-based sentiment is rarely considered. In this paper, a hybrid attention capsule network for aspect-level sentiment classification (ABASCap) was proposed. In this model, the multi-head self-attention was improved, and a context mask mechanism based on adjustable context window was proposed, so as to effectively obtain the internal association between aspects and context. Moreover, the dynamic routing algorithm and activation function in capsule network were optimized to meet the task requirements. Finally, sufficient experiments were conducted on three benchmark datasets in different domains. Compared with other baseline models, ABASCap achieved better classification results, and outperformed the state-of-the-art methods in this task after incorporating pre-training BERT.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.51 no.1
• /
• pp.8-17
• /
• 2002

This paper suggests that the immune algorithm can effectively be used in tuning of a PID controller. The artificial immune network always has a new parallel decentralized processing mechanism for various situations, since antibodies communicate to each other among different species of antibodies/B-cells through the stimulation and suppression chains among antibodies that form a large-scaled network. In addition to that, the structure of the network is not fixed, but varies continuously. That is, the artificial immune network flexibly self-organizes according to dynamic changes of external environment (meta-dynamics function). However, up to the present time, models based on the conventional crisp approach have been used to describe dynamic model relationship between antibody and antigen. Therefore, there are some problems with a less flexible result to the external behavior. On the other hand, a number of tuning technologies have been considered for the tuning of a PID controller. As a less common method, the fuzzy and neural network or its combined techniques are applied. However, in the case of the latter, yet, it is not applied in the practical field, in the former, a higher experience and technology is required during tuning procedure. In addition to that, tuning performance cannot be guaranteed with regards to a plant with non-linear characteristics or many kinds of disturbances. Along with these, this paper used immune algorithm in order that a PID controller can be more adaptable controlled against the external condition, including moise or disturbance of plant. Parameters P, I, D encoded in antibody randomly are allocated during selection processes to obtain an optimal gain required for plant. The result of study shows the artificial immune can effectively be used to tune, since it can more fit modes or parameters of the PID controller than that of the conventional tuning methods.

• Proceedings of the KAIS Fall Conference
• /
• 2003.11a
• /
• pp.141-148
• /
• 2003

Customer retention is a common concern for many industries and a critical issue for the survival in today's greatly compressed marketplace. Current customer retention models only focus on detection of potential defectors based on the likelihood of defection by using demographic and customer profile information. In this paper, we propose a dynamic procedure for defection detection and prevention using past and current customer behavior by utilizing SOM and Markov chain. The basic idea originates from the observation that a customer has a tendency to change his behavior (i.e. trim-out his usage volumes) before his eventual withdrawal. This gradual pulling out process offers the company the opportunity to detect the defection signals. With this approach, we have two significant benefits compared with existing defection detection studies. First, our procedure can predict when the potential defectors could withdraw and this feature helps to give marketing managers ample lead-time for preparing defection prevention plans. The second benefit is that our approach can provide a procedure for not only defection detection but also defection prevention, which could suggest the desirable behavior state for the next period so as to lower the likelihood of defection. We applied our dynamic procedure for defection detection and prevention to the online gaming industry. Our suggested procedure could predict potential defectors without deterioration of prediction accuracy compared to that of the MLP neural network and DT.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.7
• /
• pp.1749-1758
• /
• 1997

This paper suggests the development of dynamic forecasting model for short-term power demand based on Radial Basis Function Network and Pal's GLVQ algorithm. Radial Basis Function methods are often compared with the backpropagation training, feed-forward network, which is the most widely used neural network paradigm. The Radial Basis Function Network is a single hidden layer feed-forward neural network. Each node of the hidden layer has a parameter vector called center. This center is determined by clustering algorithm. Theatments of classical approached to clustering methods include theories by Hartigan(K-means algorithm), Kohonen(Self Organized Feature Maps %3A SOFM and Learning Vector Quantization %3A LVQ model), Carpenter and Grossberg(ART-2 model). In this model, the first approach organizes the load pattern into two clusters by Pal's GLVQ clustering algorithm. The reason of using GLVQ algorithm in this model is that GLVQ algorithm can classify the patterns better than other algorithms. And the second approach forecasts hourly load patterns by radial basis function network which has been constructed two hidden nodes. These nodes are determined from the cluster centers of the GLVQ in first step. This model was applied to forecast the hourly loads on Mar. $4^{th},\;Jun.\;4^{th},\;Jul.\;4^{th},\;Sep.\;4^{th},\;Nov.\;4^{th},$ 1995, after having trained the data for the days from Mar. $1^{th}\;to\;3^{th},\;from\;Jun.\;1^{th}\;to\;3^{th},\;from\;Jul.\;1^{th}\;to\;3^{th},\;from\;Sep.\;1^{th}\;to\;3^{th},\;and\;from\;Nov.\;1^{th}\;to\;3^{th},$ 1995, respectively. In the experiments, the average absolute errors of one-hour ahead forecasts on utility actual data are shown to be 1.3795%.

• Proceedings of the KIEE Conference
• /
• 1997.07c
• /
• pp.1021-1024
• /
• 1997

This paper has presented a method for self-tuning tile PID controller using a BP method of multilayered NNs. The proposed controller employ input signal as a learning signal of PID control. The proposed controller is applied to load-frequency control of power system and it is investigated a dynamic characteristic. The simulation results shows that proposed NN STPID controller has the good dynamics responses against load disturbances.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.10 no.6
• /
• pp.616-628
• /
• 2000

In this paper, we present the modeling of the ozone prediction system using Neuro-Fuzzy approaches. The mechanism of ozone concentration is highly complex, nonlinear, and nonstationary, the modeling of ozone prediction system has many problems and the results of prediction is not a good performance so far. The Dynamic Polynomial Neural Network(DPNN) which employs a typical algorithm of GMDH(Group Method of Data Handling) is a useful method for data analysis, identification of nonlinear complex system, and prediction of a dynamical system. The structure of the final model is compact and the computation speed to produce an output is faster than other modeling methods. In addition to DPNN, this paper also includes a Fuzzy Logic Method for modeling of ozone prediction system. The results of each modeling method and the performance of ozone prediction are presented. The proposed method shows that the prediction to the ozone concentration based upon Neuro-Fuzzy approaches gives us a good performance for ozone prediction in high and low ozone concentration with the ability of superior data approximation and self organization.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.3
• /
• pp.35-42
• /
• 2021

In this paper, we propose a deep learning-based person re-identification method using a three-dimensional RGB-Depth Xtion2 camera considering joint coordinates and dynamic features(velocity, acceleration). The main idea of the proposed identification methodology is to easily extract gait data such as joint coordinates, dynamic features with an RGB-D camera and automatically identify gait patterns through a self-designed one-dimensional convolutional neural network classifier(1D-ConvNet). The accuracy was measured based on the F1 Score, and the influence was measured by comparing the accuracy with the classifier model (JC) that did not consider dynamic characteristics. As a result, our proposed classifier model in the case of considering the dynamic characteristics(JCSpeed) showed about 8% higher F1-Score than JC.

• Smart Media Journal
• /
• v.12 no.5
• /
• pp.28-35
• /
• 2023

In this study, we propose a new light-weight model RoutingConvNet with fewer parameters to improve the applicability and practicality of speech emotion recognition. To reduce the number of learnable parameters, the proposed model connects bidirectional MFCCs on a channel-by-channel basis to learn long-term emotion dependence and extract contextual features. A light-weight deep CNN is constructed for low-level feature extraction, and self-attention is used to obtain information about channel and spatial signals in speech signals. In addition, we apply dynamic routing to improve the accuracy and construct a model that is robust to feature variations. The proposed model shows parameter reduction and accuracy improvement in the overall experiments of speech emotion datasets (EMO-DB, RAVDESS, and IEMOCAP), achieving 87.86%, 83.44%, and 66.06% accuracy respectively with about 156,000 parameters. In this study, we proposed a metric to calculate the trade-off between the number of parameters and accuracy for performance evaluation against light-weight.