• Information Systems Review
• /
• v.23 no.1
• /
• pp.155-172
• /
• 2021

With the recent increase in online product purchases, a recommender system that recommends products considering users' preferences has still been studied. The recommender system provides personalized product recommendation services to users. Collaborative Filtering (CF) using user ratings on products is one of the most widely used recommendation algorithms. During CF, the item-based method identifies the user's product by using ratings left on the product purchased by the user and obtains the similarity between the purchased product and the unpurchased product. CF takes a lot of time to calculate the similarity between products. In particular, it takes more time when using text-based big data such as review data of Amazon store. This paper suggests a hybrid recommendation system using a 2-phase methodology and text data mining to calculate the similarity between products easily and quickly. To this end, we collected about 980,000 online consumer ratings and review data from the online commerce store, Amazon Kinder Store. As a result of several experiments, it was confirmed that the suggested hybrid recommendation system reflecting the user's rating and review data has resulted in similar recommendation time, but higher accuracy compared to the CF-based benchmark recommender systems. Therefore, the suggested system is expected to increase the user's satisfaction and increase its sales.

• Journal of the Korean Crystal Growth and Crystal Technology
• /
• v.34 no.2
• /
• pp.73-77
• /
• 2024

The spinel-type oxide (Ni_x, Mn_y, Co_3-x-y)O₄ (NMC) is widely utilized as a material for temperature sensors with a negative temperature coefficient (NTC), finding applications across various industries including electric vehicle battery management systems. Typically, NMC is manufactured using solid-state reaction methods employing powders of Ni, Mn, and Co compounds, with the densification process through sintering recognized as a crucial factor determining the electrical properties of the temperature sensor material. In this study, NMC pellets were synthesized via solid-state reaction and their crystallographic and microstructural characteristics were investigated. Also, the activation energy for densification behavior during the sintering process was determined. According to the analysis results, the room temperature resistance of the NMC pellets was measured at 10.03 Kohm, with the sensitivity parameter, B-value, recorded at 3601.8 K, indicating their potential applicability as temperature sensors across various industrial fields. Furthermore, the activation energy for densification was found to be 273.3 ± 0.4 kJ/mol, providing valuable insights into the thermodynamic aspects of the sintering process of the NMC.

• Journal of the Korean Society of Radiology
• /
• v.18 no.1
• /
• pp.37-44
• /
• 2024

Even though MR can reveal excellent soft-tissue contrast and functional information, CT is also required for electron density information for accurate dose calculation in Radiotherapy. For the fusion of MRI and CT images in RT treatment planning workflow, patients are normally scanned on both MRI and CT imaging modalities. Recently deep-learning-based generations of CT images from MR images became possible owing to machine learning technology. This eliminated CT scanning work. This study implemented a CycleGan deep-learning-based CT image generation from MR images. Three CT generators whose learning is based on T1- , T2- , or T1-&T2-weighted MR images were created, respectively. We found that the T1-weighted MR image-based generator can generate better than other CT generators when T1-weighted MR images are input. In contrast, a T2-weighted MR image-based generator can generate better than other CT generators do when T2-weighted MR images are input. The results say that the CT generator from MR images is just outside the practical clinics and the specific weight MR image-based machine-learning generator can generate better CT images than other sequence MR image-based generators do.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.5
• /
• pp.213-221
• /
• 2006

Phone segmentation of speech waveform is especially important for concatenative text to speech synthesis which uses segmented corpora for the construction of synthetic units. because the quality of synthesized speech depends critically on the accuracy of the segmentation. In the beginning. the phone segmentation was manually performed. but it brings the huge effort and the large time delay. HMM-based approaches adopted from automatic speech recognition are most widely used for automatic segmentation in speech synthesis, providing a consistent and accurate phone labeling scheme. Even the HMM-based approach has been successful, it may locate a phone boundary at a different position than expected. In this paper. we categorized adjacent phoneme pairs and analyzed the mismatches between hand-labeled transcriptions and HMM-based labels. Then we described the dominant error patterns that must be improved for the speech synthesis. For the experiment. hand labeled standard Korean speech DB from ETRI was used as a reference DB. Time difference larger than 20ms between hand-labeled phoneme boundary and auto-aligned boundary is treated as an automatic segmentation error. Our experimental results from female speaker revealed that plosive-vowel, affricate-vowel and vowel-liquid pairs showed high accuracies, 99%, 99.5% and 99% respectively. But stop-nasal, stop-liquid and nasal-liquid pairs showed very low accuracies, 45%, 50% and 55%. And these from male speaker revealed similar tendency.

• Information Systems Review
• /
• v.22 no.1
• /
• pp.59-72
• /
• 2020

Artificial Intelligence is establishing itself as a familiar tool from an intractable concept. In this trend, financial sector is also looking to improve the problem of existing system which includes Fraud Detection System (FDS). It is being difficult to detect sophisticated cyber financial fraud using original rule-based FDS. This is because diversification of payment environment and increasing number of electronic financial transactions has been emerged. In order to overcome present FDS, this paper suggests 3 types of artificial intelligence models, Generative Adversarial Network (GAN), Deep Neural Network (DNN), and Convolutional Neural Network (CNN). GAN proves how data imbalance problem can be developed while DNN and CNN show how abnormal financial trading patterns can be precisely detected. In conclusion, among the experiments on this paper, WGAN has the highest improvement effects on data imbalance problem. DNN model reflects more effects on fraud classification comparatively.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.34 no.1
• /
• pp.129-141
• /
• 2024

Amidst the acceleration of digital transformation across various sectors, the financial market is increasingly focusing on the development of digital and electronic payment methods, including currency. Among these, Central Bank Digital Currencies (CBDC) are emerging as future digital currencies that could replace physical cash. They are stable, not subject to value fluctuation, and can be exchanged one-to-one with existing physical currencies. Recently, both domestic and international efforts are underway in researching and developing CBDCs. However, current CBDC systems face scalability issues such as delays in processing large transactions, response times, and network congestion. To build a universal CBDC system, it is crucial to resolve these scalability issues, including the low throughput and network overload problems inherent in existing blockchain technologies. Therefore, this study proposes a solution based on reinforcement learning for handling large-scale data in a CBDC environment, aiming to improve throughput and reduce network congestion. The proposed technology can increase throughput by more than 64 times and reduce network congestion by over 20% compared to existing systems.

• Information Systems Review
• /
• v.26 no.2
• /
• pp.45-57
• /
• 2024

This study develops a deep learning-based methodology for predicting Crude Palm Oil (CPO) prices. Palm oil is an essential resource across various industries due to its yield and economic efficiency, leading to increased industrial interest in its price volatility. While numerous studies have been conducted on palm oil price prediction, most rely on time series forecasting, which has inherent accuracy limitations. To address the main limitation of traditional methods-the absence of stationarity-this research introduces a novel model that uses the ratio of future prices to current prices as the dependent variable. This approach, inspired by return modeling in stock price predictions, demonstrates superior performance over simple price prediction. Additionally, the methodology incorporates the consideration of lag values of independent variables, a critical factor in multivariate time series forecasting, to eliminate unnecessary noise and enhance the stability of the prediction model. This research not only significantly improves the accuracy of palm oil price prediction but also offers an applicable approach for other economic forecasting issues where time series data is crucial, providing substantial value to the industry.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.34 no.5
• /
• pp.973-980
• /
• 2024

Fuzzing is an automated testing technique that generates a lot of testcases and monitors for exceptions to test a program. Recently, fuzzing research using machine learning has been actively proposed to solve various problems in the fuzzing process, but a comprehensive evaluation of fuzzing research using machine learning is lacking. In this paper, we analyze recent research that applies machine learning to scheduling techniques for fuzzing, categorizing them into reinforcement learning-based and supervised learning-based fuzzers. We evaluated the coverage performance of the analyzed machine learning-based fuzzers against real-world programs with four different file formats and bug detection performance against the LAVA-M dataset. The results showed that AFL-HIER, which applied seed clustering and seed scheduling with reinforcement learning outperformed in coverage and bug detection. In the case of supervised learning, it showed high coverage on tcpdumps with high code complexity, and its superior bug detection performance when applied to hybrid fuzzing. This research shows that performance of machine learning-based fuzzer is better when both machine learning and additional fuzzing techniques are used to optimize the fuzzing process. Future research is needed on practical and robust machine learning-based fuzzing techniques that can be effectively applied to programs that handle various input formats.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.43 no.2 s.308
• /
• pp.19-29
• /
• 2006

As the security of personal information is becoming more important in mobile phones, we are starting to apply iris recognition technology to these devices. In conventional iris recognition, magnified iris images are required. For that, it has been necessary to use large magnified zoom & focus lens camera to capture images, but due to the requirement about low size and cost of mobile phones, the zoom & focus lens are difficult to be used. However, with rapid developments and multimedia convergence trends in mobile phones, more and more companies have built mega-pixel cameras into their mobile phones. These devices make it possible to capture a magnified iris image without zoom & focus lens. Although facial images are captured far away from the user using a mega-pixel camera, the captured iris region possesses sufficient pixel information for iris recognition. However, in this case, the eye region should be detected for accurate iris recognition in facial images. So, we propose a new fast iris detection method, which is appropriate for mobile phones based on corneal specular reflection. To detect specular reflection robustly, we propose the theoretical background of estimating the size and brightness of specular reflection based on eye, camera and illuminator models. In addition, we use the successive On/Off scheme of the illuminator to detect the optical/motion blurring and sunlight effect on input image. Experimental results show that total processing time(detecting iris region) is on average 65ms on a Samsung SCH-S2300 (with 150MHz ARM 9 CPU) mobile phone. The rate of correct iris detection is 99% (about indoor images) and 98.5% (about outdoor images).

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.12 no.6
• /
• pp.657-665
• /
• 2019

In this study, we conducted a study on the transcranial magnetic electrode, a method for the study of dementia and muscle pain, a neurodegenerative disease caused by an aging society, which is becoming a problem worldwide. In particular, transcranial magnetic electrodes have been studied to improve their ability to be deteriorated by dementia symptoms such as speech, cognitive ability, and memory by outputting magnetism deep into the brain using coils on the head epidermis. In this study, simulation was performed using Maxwell 3D program for the design of coil, the core of transcranial magnetic electrode. As a result of the simulation comparison between the coil designed by the previous research and the coil through the research and development, the output was found to be superior to the conventional designed coil. The graphs of the coil outputs of B-Field and H-Field are found to be symmetrical, but the symmetry between each coil is pseudo-symmetrical and not accurate. Based on these results, an experiment was conducted to confirm whether the output of the head epidermis through both coils is possible. In the magnitude field of the reverse-coil 2-coil analysis, the maximum output was 3.3920e + 004 H [A_per_meter], and the vector field showed the strongest magnetic field around 35 to 165 degrees. It was confirmed that the magnetic output canceled due to the magnetic output. In the case of the forward 2-coil, a maximum of 3.2348e + 004H [A_per_meter] similar to the reverse coil was observed, but in the case of the vector field, the magnetic output regarding the forward output and the head skin output was confirmed. However, when the height change in the output coil, the magnetic output was reduced.