• Journal of Korea Multimedia Society
• /
• v.24 no.3
• /
• pp.382-395
• /
• 2021

In this study, character recognition using deep learning is performed among the various defects in the PCB, the purpose of which is to check whether the printed characters are printed correctly on top of components, or the incorrect parts are attached. Generally, character recognition may be perceived as not a difficult problem when considering MNIST, but the printed letters on the PCB component data are difficult to collect, and have very high redundancy. So if a deep learning model is trained with original data without any preprocessing, it can lead to over fitting problems. Therefore, this study aims to reduce the redundancy to the smallest dataset that can represent large amounts of data collected in limited production sites, and to create datasets through data enhancement to train a flexible deep learning model can be used in various production sites. Moreover, ResNet model verifies to determine which combination of datasets is the most effective. This study discusses how to reduce and augment data that is constantly occurring in real PCB production lines, and discusses how to select coresets to learn and apply deep learning models in real sites.

• Clean Technology
• /
• v.28 no.2
• /
• pp.138-146
• /
• 2022

Electricity has become a factor that dramatically affects the market economy. The day-ahead system marginal price determines electricity prices, and system marginal price forecasting is critical in maintaining energy management systems. There have been several studies using mathematics and machine learning models to forecast the system marginal price, but few studies have been conducted to develop, compare, and analyze various machine learning and deep learning models based on a data-driven framework. Therefore, in this study, different machine learning algorithms (i.e., autoregressive-based models such as the autoregressive integrated moving average model) and deep learning networks (i.e., recurrent neural network-based models such as the long short-term memory and gated recurrent unit model) are considered and integrated evaluation metrics including a forecasting test and information criteria are proposed to discern the optimal forecasting model. A case study of South Korea using long-term time-series system marginal price data from 2016 to 2021 was applied to the developed framework. The results of the study indicate that the autoregressive integrated moving average model (R-squared score: 0.97) and the gated recurrent unit model (R-squared score: 0.94) are appropriate for system marginal price forecasting. This study is expected to contribute significantly to energy management systems and the suggested framework can be explicitly applied for renewable energy networks.

• Research in Mathematical Education
• /
• v.27 no.1
• /
• pp.129-150
• /
• 2024

This study utilized South Korean elementary and middle school student data to examine the longitudinal change trajectories of learning motivation types according to the longitudinal change trajectories of mathematics academic achievement. Growth mixture modeling, latent growth model, and multiple indicator latent growth model were used to examine various change trajectories for longitudinal data. As a result of the analysis, it was classified into 4 subgroups with similar longitudinal change trajectories of mathematics academic achievement, and the characteristics of the mathematics subject, which emphasize systematicity, appeared. Furthermore, higher mathematics academic achievement was associated with higher self-determination and higher academic motivation. And as the grade level increases, amotivation increases and self-determination decreases. This study suggests that teaching and learning support using this is necessary because the level of learning motivation according to self-determination is different depending on the level of mathematics academic achievement reflecting the characteristics of the student.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.19 no.4
• /
• pp.991-1000
• /
• 2015

This paper is studying for New Online Learning Model on Smart digital era. It can expect the result of learning degree on Online Learning Class. Using 7-factors of Online Class's operating policy, make the expectation model by 'decision tree' method. And through applying another class, we can getting a generality. Finally, Over the traditional Online Class model, we can take the real-time status of Online class learning degree. It is useful both students and teacher. It is the one of 'Smart learning Model'.

• Journal of Information Technology Services
• /
• v.21 no.3
• /
• pp.63-72
• /
• 2022

The extinction crisis of local cities, caused by a population density increase phenomenon in capital regions, directly causes the increase of vacant houses in local cities. According to population and housing census, Gunsan-si has continuously shown increasing trend of vacant houses during 2015 to 2019. In particular, since Gunsan-si is the city which suffers from doughnut effect and industrial decline, problems regrading to vacant house seems to exacerbate. This study aims to provide a foundation of a system which can predict and deal with the building that has high risk of becoming vacant house through implementing a data driven vacant house prediction machine learning model. Methodologically, this study analyzes three types of machine learning model by differing the data components. First model is trained based on building register, individual declared land value, house price and socioeconomic data and second model is trained with the same data as first model but with additional POI(Point of Interest) data. Finally, third model is trained with same data as the second model but with excluding water usage and electricity usage data. As a result, second model shows the best performance based on F1-score. Random Forest, Gradient Boosting Machine, XGBoost and LightGBM which are tree ensemble series, show the best performance as a whole. Additionally, the complexity of the model can be reduced through eliminating independent variables that have correlation coefficient between the variables and vacant house status lower than the 0.1 based on absolute value. Finally, this study suggests XGBoost and LightGBM based machine learning model, which can handle missing values, as final vacant house prediction model.

• The Journal of Korea Robotics Society
• /
• v.15 no.4
• /
• pp.398-403
• /
• 2020

Robot's throwing control is difficult to accurately calculate because of air resistance and rotational inertia, etc. This complexity can be solved by using machine learning. Reinforcement learning using reward function puts limit on adapting to new environment for robots. Therefore, this paper applied deep reinforcement learning using neural network without reward function. Throwing is evaluated as a success or failure. AI network learns by taking the target position and control policy as input and yielding the evaluation as output. Then, the task is carried out by predicting the success probability according to the target location and control policy and searching the policy with the highest probability. Repeating this task can result in performance improvements as data accumulates. And this model can even predict tasks that were not previously attempted which means it is an universally applicable learning model for any new environment. According to the data results from 520 experiments, this learning model guarantees 75% success rate.

• Journal of the Korean Society of Safety
• /
• v.38 no.4
• /
• pp.1-7
• /
• 2023

The steel wheels of urban railway vehicles gather a lot of data through regular measurements during maintenance. However, limited research has been carried out utilizing this data, resulting in difficulties predicting the maintenance period. This paper studied a machine learning model suitable for mileage prediction by studying the characteristics of mileage change according to diameter and flange thickness changes. The results of this study indicate that the larger the diameter, the longer the travel distance, and the longest flange thickness is at 30 mm, which gradually shortened at other times. As a result of research on the machine learning prediction model, it was confirmed that the random forest model is the optimal model with a high coefficient of determination and a low root mean square error.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• v.27 no.2
• /
• pp.135-145
• /
• 2023

Recently, as deep learning technology has developed, various deep learning technologies have been introduced in handwritten recognition, greatly contributing to performance improvement. The recognition accuracy of handwritten Hangeul recognition has also improved significantly, but prior research has focused on recognizing 520 Hangul characters or 2,350 Hangul characters using SERI95 data or PE92 data. In the past, most of the expressions were possible with 2,350 Hangul characters, but as globalization progresses and information and communication technology develops, there are many cases where various foreign words need to be expressed in Hangul. In this paper, we propose a model that recognizes and combines the consonants, medial vowels, and final consonants of a Korean syllable using a multi-label classification model, and achieves a high recognition accuracy of 98.38% as a result of learning with the public data of Korean handwritten characters, PE92. In addition, this model learned only 2,350 Hangul characters, but can recognize the characters which is not included in the 2,350 Hangul characters

• IEMEK Journal of Embedded Systems and Applications
• /
• v.19 no.3
• /
• pp.151-158
• /
• 2024

Although on-device artificial intelligence (AI) has gained attention to diagnosing machine faults in real time, most previous studies did not consider the model retraining and redeployment processes that must be performed in real-world industrial environments. Our study addresses this challenge by proposing an on-device AI-based real-time machine fault diagnosis system that utilizes continual learning. Our proposed system includes a lightweight convolutional neural network (CNN) model, a continual learning algorithm, and a real-time monitoring service. First, we developed a lightweight 1D CNN model to reduce the cost of model deployment and enable real-time inference on the target edge device with limited computing resources. We then compared the performance of five continual learning algorithms with three public bearing fault datasets and selected the most effective algorithm for our system. Finally, we implemented a real-time monitoring service using an open-source data visualization framework. In the performance comparison results between continual learning algorithms, we found that the replay-based algorithms outperformed the regularization-based algorithms, and the experience replay (ER) algorithm had the best diagnostic accuracy. We further tuned the number and length of data samples used for a memory buffer of the ER algorithm to maximize its performance. We confirmed that the performance of the ER algorithm becomes higher when a longer data length is used. Consequently, the proposed system showed an accuracy of 98.7%, while only 16.5% of the previous data was stored in memory buffer. Our lightweight CNN model was also able to diagnose a fault type of one data sample within 3.76 ms on the Raspberry Pi 4B device.

• International Journal of Advanced Culture Technology
• /
• v.10 no.1
• /
• pp.160-169
• /
• 2022

This qualitative study aimed to investigate ways to improve effective cooperative learning from students' perspective by understanding and analyzing the learning experiences of nursing students who participated in a collaborative problem-solving learning model. Data were collected through focus group interviews and reflection journals of six second-year nursing students from G-university in J-city who participated in a collaborative problem-solving learning model course. The interview data were analyzed and divided into 3 categories and 10 subcategories according to the six-step thematic analysis method proposed by Braun and Clarke. The results of analyzing the interviews were considered based on three areas: preparation before learning, the process of collaborating as a cooperative learning experience, and solutions and expectations after learning. The participants felt frustrated because collaborative problem-solving took more time for individual learning than traditional methods did and would not allow them to check the correct answers immediately. However, they gained new experiences by solving problems and engaging in discussions within their learning community. The participants' expectations included material that could help their learning, measures to prevent free-riders, and consideration of the learning process in evaluation factors. Although this study has sample limitations by targeting nursing students in only one region, it can be used to help operate collaborative problem-solving classes, as it reflects the real experiences and opinions of students.