• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.47 no.2
• /
• pp.155-167
• /
• 2024

Distribution and logistics industries contribute some of the biggest GDP(gross domestic product) in South Korea and the number of related companies are quarter of the total number of industries in the country. The number of retail tech companies are quickly increased due to the acceleration of the online and untact shopping trend. Furthermore, major distribution and logistics companies try to achieve integrated data management with the fulfillment process. In contrast, small and medium distribution companies still lack of the capacity and ability to develop digital innovation and smartization. Therefore, in this paper, a deep learning-based demand forecasting & recommendation model is proposed to improve business competitiveness. The proposed model is developed based on real sales transaction data to predict future demand for each product. The proposed model consists of six deep learning models, which are MLP(multi-layers perception), CNN(convolution neural network), RNN(recurrent neural network), LSTM(long short term memory), Conv1D-BiLSTM(convolution-long short term memory) for demand forecasting and collaborative filtering for the recommendation. Each model provides the best prediction result for each product and recommendation model can recommend best sales product among companies own sales list as well as competitor's item list. The proposed demand forecasting model is expected to improve the competitiveness of the small and medium-sized distribution and logistics industry.

• Journal of Korea Society of Industrial Information Systems
• /
• v.29 no.4
• /
• pp.13-24
• /
• 2024

In this paper, we analyzed the performance of normalization according to various neural network models and input-output structures. For the analysis, a simulation-based dataset for noise environments with homogeneous and up to three interfering signals was used. As a result, the end-to-end structure that directly outputs noise variance showed superior performance when using a 1-D convolutional neural network and BiLSTM model, and was analyzed to be particularly robust against interference signals. This is because the 1-D convolutional neural network and bidirectional long short-term memory models have stronger inductive bias than the multilayer perceptron and transformer models. The analysis of this paper are expected to be used as a useful reference for future research on deep learning-based normalization.

• Korean Journal of Remote Sensing
• /
• v.37 no.5_1
• /
• pp.1149-1161
• /
• 2021

With a trend of the utilization of computer vision for satellite images, cloud detection using deep learning also attracts attention recently. In this study, we conducted a U-Net cloud detection modeling using SPARCS (Spatial Procedures for Automated Removal of Cloud and Shadow) Cloud Dataset with the image data augmentation and carried out 10-fold cross-validation for an objective assessment of the model. Asthe result of the blind test for 1800 datasets with 512 by 512 pixels, relatively high performance with the accuracy of 0.821, the precision of 0.847, the recall of 0.821, the F1-score of 0.831, and the IoU (Intersection over Union) of 0.723. Although 14.5% of actual cloud shadows were misclassified as land, and 19.7% of actual clouds were misidentified as land, this can be overcome by increasing the quality and quantity of label datasets. Moreover, a state-of-the-art DeepLab V3+ model and the NAS (Neural Architecture Search) optimization technique can help the cloud detection for CAS500 (Compact Advanced Satellite 500) in South Korea.

• Journal of Multimedia Information System
• /
• v.6 no.4
• /
• pp.209-216
• /
• 2019

Over the past decade, researchers were able to solve complex medical problems as well as acquire deeper understanding of entire issue due to the availability of machine learning techniques, particularly predictive algorithms and automatic recognition of patterns in medical imaging. In this study, a technique called transfer learning has been utilized to classify Magnetic Resonance (MR) images by a pre-trained Convolutional Neural Network (CNN). Rather than training an entire model from scratch, transfer learning approach uses the CNN model by fine-tuning them, to classify MR images into Alzheimer's disease (AD), mild cognitive impairment (MCI) and normal control (NC). The performance of this method has been evaluated over Alzheimer's Disease Neuroimaging (ADNI) dataset by changing the learning rate of the model. Moreover, in this study, in order to demonstrate the transfer learning approach we utilize different pre-trained deep learning models such as GoogLeNet, VGG-16, AlexNet and ResNet-18, and compare their efficiency to classify AD. The overall classification accuracy resulted by GoogLeNet for training and testing was 99.84% and 98.25% respectively, which was exceptionally more than other models training and testing accuracies.

• International Journal of Naval Architecture and Ocean Engineering
• /
• v.12 no.1
• /
• pp.428-439
• /
• 2020

The continuous development of information and communication technologies has resulted in an exponential increase in data. Consequently, technologies related to data analysis are growing in importance. The shipbuilding industry has high production uncertainty and variability, which has created an urgent need for data analysis techniques, such as machine learning. In particular, the industry cannot effectively respond to changes in the production-related standard time information systems, such as the basic cycle time and lead time. Improvement measures are necessary to enable the industry to respond swiftly to changes in the production environment. In this study, the lead times for fabrication, assembly of ship block, spool fabrication and painting were predicted using machine learning technology to propose a new management method for the process lead time using a master data system for the time element in the production data. Data preprocessing was performed in various ways using R and Python, which are open source programming languages, and process variables were selected considering their relationships with the lead time through correlation analysis and analysis of variables. Various machine learning, deep learning, and ensemble learning algorithms were applied to create the lead time prediction models. In addition, the applicability of the proposed machine learning methodology to standard work hour prediction was verified by evaluating the prediction models using the evaluation criteria, such as the Mean Absolute Percentage Error (MAPE) and Root Mean Squared Logarithmic Error (RMSLE).

• Journal of Internet Computing and Services
• /
• v.21 no.3
• /
• pp.133-144
• /
• 2020

The number of elderly people with dementia is increasing as fast as the proportion of older people due to aging, which creates a social and economic burden. In particular, dementia care costs, including indirect costs such as increased care costs due to lost caregiver hours and caregivers, have grown exponentially over the years. In order to reduce these costs, it is urgent to introduce a management system to care for dementia patients. Therefore, this study proposes a sensor-based abnormal behavior detection system to manage dementia patients who live alone or in an environment where they cannot always take care of dementia patients. Existing studies were merely evaluating behavior or evaluating normal behavior, and there were studies that perceived behavior by processing images, not data from sensors. In this study, we recognized the limitation of real data collection and used both the auto-encoder, the unsupervised learning model, and the LSTM, the supervised learning model. Autoencoder, an unsupervised learning model, trained normal behavioral data to learn patterns for normal behavior, and LSTM further refined classification by learning behaviors that could be perceived by sensors. The test results show that each model has about 96% and 98% accuracy and is designed to pass the LSTM model when the autoencoder outlier has more than 3%. The system is expected to effectively manage the elderly and dementia patients who live alone and reduce the cost of caring.

• Journal of The Korean Association of Information Education
• /
• v.24 no.6
• /
• pp.539-550
• /
• 2020

This study establishes the scientific basis for the use of digital learning materials through the education-neurological research method and derives implications for education based on education-neurological understandings. The main findings of the education-neurological analysis of digital learning materials are as follows: First, various sensory stimuli go through multiple sensory neurons and deep sections of the upper sphere and make possible the cooperative processing of information. Second, indirect experience from digital learning materials helps students understand the learning contents vividly through the mirror neuron system. Third, positive emotions originating from digital learning materials promote functions of dopamine, the reticular activating system, frontal-striatal circuit, cerebrum cortex. Based on the findings, the study suggests the following educational implications. First of all, when selecting digital learning materials, teachers should consider expression forms, learning contents, the flow of classes, and the adverse effects of digital learning materials. Next, it is effective to utilize digital learning materials in the lecture for provoking curiosity and enjoyment, maintaining interest and effort, and reviewing what students learned.

• Journal of Information Processing Systems
• /
• v.17 no.6
• /
• pp.1071-1082
• /
• 2021

Since the smart factory has been recently recognized as an industrial core requirement, various mechanisms to ensure efficient and stable operation have attracted much attention. This attention is based on the fact that in a smart factory environment where operating processes, such as facility control, data collection, and decision making are automated, the disruption of processes due to problems such as facility anomalies causes considerable losses. Although many studies have considered methods to prevent such losses, few have investigated how to effectively apply the solutions. This study proposes a Kubernetes based system applied in a smart factory providing effective operation and facility management. To develop the system, we employed a useful and popular open source project, and adopted deep learning based anomaly detection model for multi-sensor anomaly detection. This can be easily modified without interruption by changing the container image for inference. Through experiments, we have verified that the proposed method can provide system stability through nondisruptive maintenance, monitoring and non-disruptive updates for anomaly detection models.

• Journal of the Korean Society of Systems Engineering
• /
• v.17 no.2
• /
• pp.91-97
• /
• 2021

User interface for cancer classification system is a software application with clinician's friendly tools and functions to diagnose cancer from pathology images. Pathology evolved from manual diagnosis to computer-aided diagnosis with the help of Artificial Intelligence tools and algorithms. In this paper, we explained each block of the project life cycle for the implementation of automated breast cancer classification software using AI and machine learning algorithms to classify normal and invasive breast histology images. The system was designed to help the pathologists in an automatic and efficient diagnosis of breast cancer. To design the classification model, Hematoxylin and Eosin (H&E) stained breast histology images were obtained from the ICIAR Breast Cancer challenge. These images are stain normalized to minimize the error that can occur during model training due to pathological stains. The normalized dataset was fed into the ResNet-34 for the classification of normal and invasive breast cancer images. ResNet-34 gave 94% accuracy, 93% F Score, 95% of model Recall, and 91% precision.

• Annual Conference of KIPS
• /
• 2021.11a
• /
• pp.446-448
• /
• 2021

In training a deep learning model, it is crucial to tune various hyperparameters and gain speed and accuracy. While hyperparameters that mathematically induce convergence impact training speed, system parameters that affect host-to-device transfer are also crucial. Therefore, it is important to properly tune and select parameters that influence the data loader as a system parameter in overall time acceleration. We propose an automated framework called Num Worker Tuner (NWT) to address this problem. This method finds the appropriate number of multi-processing subprocesses through the search space and accelerates the learning through the number of subprocesses. Furthermore, this method allows memory efficiency and speed-up by tuning the system-dependent parameter, the number of multi-process spawns.