• 대한원격탐사학회지
• /
• 제39권1호
• /
• pp.87-97
• /
• 2023

Unmanned aerial vehicles (UAVs) can capture high-resolution imagery from a variety of viewing angles and altitudes; they are generally limited to collecting images of small scenes from larger regions. To improve the utility of UAV-appropriated datasetsfor use with deep learning applications, multiple datasets created from variousregions under different conditions are needed. To demonstrate a powerful new method for integrating heterogeneous UAV datasets, this paper applies a combined segmentation network (CSN) to share UAVid and semantic drone dataset encoding blocks to learn their general features, whereas its decoding blocks are trained separately on each dataset. Experimental results show that our CSN improves the accuracy of specific classes (e.g., cars), which currently comprise a low ratio in both datasets. From this result, it is expected that the range of UAV dataset utilization will increase.

• 생물정신의학
• /
• 제30권1호
• /
• pp.24-30
• /
• 2023

Objectives Alzheimer's disease (AD) is the most common form of dementia in older adults, damaging the brain and resulting in impaired memory, thinking, and behavior. The identification of differentially expressed genes and related pathways among affected brain regions can provide more information on the mechanisms of AD. The aim of our study was to identify differentially expressed genes associated with AD and combined biomarkers among them to improve AD risk prediction accuracy. Methods Machine learning methods were used to compare the performance of the identified combined biomarkers. In this study, three publicly available gene expression datasets from the hippocampal brain region were used. Results We detected 31 significant common genes from two different microarray datasets using the limma package. Some of them belonged to 11 biological pathways. Combined biomarkers were identified in two microarray datasets and were evaluated in a different dataset. The performance of the predictive models using the combined biomarkers was superior to those of models using a single gene. When two genes were combined, the most predictive gene set in the evaluation dataset was ATR and PRKCB when linear discriminant analysis was applied. Conclusions Combined biomarkers showed good performance in predicting the risk of AD. The constructed predictive nomogram using combined biomarkers could easily be used by clinicians to identify high-risk individuals so that more efficient trials could be designed to reduce the incidence of AD.

• 한국컴퓨터정보학회논문지
• /
• 제29권4호
• /
• pp.39-46
• /
• 2024

많은 연구자들이 다양한 모델을 이용하여 물의 수질을 평가하기 위해 노력하고 있다. 평가 모델에는 결측값이 없는 데이터셋이 필요하지만, 관측 데이터셋에는 결측값이 다수 포함되는 것이 현실이다. 단순히 결측값을 삭제하는 방법은 경우에 따라 기저 데이터의 분포를 왜곡시키고 모델의 예측성능에도 편의(bias)를 불러올 위험성이 있다. 본 연구에서는 수질 데이터의 결측값 처리에 적합한 기법을 탐색하기 위해, 기존의 KNN과 MICE Imputation, 그리고 생성형 신경망 모델인 Autoencoder와 Denoising Autoencoder를 기반으로 몇 가지 대치 기법을 실험하였다. 실험 결과, KNN과 MICE Imputation의 결과를 평균한 Combined Imputation이 실측치에 가장 가깝게 값을 추정하였으며, 이 기법을 적용하여 결측값을 처리한 관측 데이터셋을 support vector machine과 ensemble 기반의 분류 모델로 평가한 결과, 결측값을 삭제했을 때에 비해 Accuracy, F1 score, ROC-AUC score, 그리고 MCC(Mathews Correlation Coefficient) 지표가 향상되었다.

• Journal of information and communication convergence engineering
• /
• 제21권2호
• /
• pp.167-173
• /
• 2023

Hangul is unique compared to other Asian languages because of its simple letter forms that combine to create syllabic shapes. There are 24 basic letters that can be combined to form 27 additional complex letters. This produces 51 graphemes. Hangul optical character recognition has been a research topic for some time; however, handwritten Hangul recognition continues to be challenging owing to the various writing styles, slants, and cursive-like nature of the handwriting. In this study, a dataset containing thousands of samples of 51 Hangul graphemes was gathered from 110 freshmen university students to create a robust dataset with high variance for training an artificial neural network. The collected dataset included 2200 samples for each consonant grapheme and 1100 samples for each vowel grapheme. The dataset was normalized to the MNIST digits dataset, trained in three neural networks, and the obtained results were compared.

• 대한임베디드공학회논문지
• /
• 제15권4호
• /
• pp.197-204
• /
• 2020

The ImageNet dataset is a large scale dataset and contains various natural scene images. In this paper, we propose a convolutional neural network (CNN)-based weighted ensemble technique for the ImageNet classification task. First, in order to fuse several models, our technique uses weights for each model, unlike the existing average-based ensemble technique. Then we propose an algorithm that automatically finds the coefficients used in later ensemble process. Our algorithm sequentially selects the model with the best performance of the validation set, and then obtains a weight that improves performance when combined with existing selected models. We applied the proposed algorithm to a total of 13 heterogeneous models, and as a result, 5 models were selected. These selected models were combined with weights, and we achieved 3.297% Top-5 error rate on the ImageNet test dataset.

• International Journal of Contents
• /
• 제17권4호
• /
• pp.1-15
• /
• 2021

Named entity recognition (NER) is a crucial task for NLP, which aims to extract information from texts. To build NER systems, deep learning (DL) models are learned with dictionary features by mapping each word in the dataset to dictionary features and generating a unique index. However, this technique might generate noisy labels, which pose significant challenges for the NER task. In this paper, we proposed DL-dictionary features, and evaluated them on two datasets, including the OntoNotes 5.0 dataset and our new infectious disease outbreak dataset named GFID. We used (1) a Bidirectional Long Short-Term Memory (BiLSTM) character and (2) pre-trained embedding to concatenate with (3) our proposed features, named the Convolutional Neural Network (CNN), BiLSTM, and self-attention dictionaries, respectively. The combined features (1-3) were fed through BiLSTM - Conditional Random Field (CRF) to predict named entity classes as outputs. We compared these outputs with other predictions of the BiLSTM character, pre-trained embedding, and dictionary features from previous research, which used the exact matching and partial matching dictionary technique. The findings showed that the model employing our dictionary features outperformed other models that used existing dictionary features. We also computed the F1 score with the GFID dataset to apply this technique to extract medical or healthcare information.

• International Journal of Computer Science & Network Security
• /
• 제22권2호
• /
• pp.1-8
• /
• 2022

This work provides a reliable and classified stocks dataset merged with Saudi stock news. This dataset allows researchers to analyze and better understand the realities, impacts, and relationships between stock news and stock fluctuations. The data were collected from the Saudi stock market via the Corporate News (CN) and Historical Data Stocks (HDS) datasets. As their names suggest, CN contains news, and HDS provides information concerning how stock values change over time. Both datasets cover the period from 2011 to 2019, have 30,098 rows, and have 16 variables-four of which they share and 12 of which differ. Therefore, the combined dataset presented here includes 30,098 published news pieces and information about stock fluctuations across nine years. Stock news polarity has been interpreted in various ways by native Arabic speakers associated with the stock domain. Therefore, this polarity was categorized manually based on Arabic semantics. As the Saudi stock market massively contributes to the international economy, this dataset is essential for stock investors and analyzers. The dataset has been prepared for educational and scientific purposes, motivated by the scarcity of data describing the impact of Saudi stock news on stock activities. It will, therefore, be useful across many sectors, including stock market analytics, data mining, statistics, machine learning, and deep learning. The data evaluation is applied by testing the data distribution of the categories and the sentiment prediction-the data distribution over classes and sentiment prediction accuracy. The results show that the data distribution of the polarity over sectors is considered a balanced distribution. The NB model is developed to evaluate the data quality based on sentiment classification, proving the data reliability by achieving 68% accuracy. So, the data evaluation results ensure dataset reliability, readiness, and high quality for any usage.

• 한국컴퓨터정보학회논문지
• /
• 제21권11호
• /
• pp.121-126
• /
• 2016

In recommender systems based on collaborative filtering, measuring similarity is very critical for determining the range of recommenders. Data sparsity problem is fundamental in collaborative filtering systems, which is partly solved by Jaccard coefficient combined with traditional similarity measures. This study proposes a new coefficient for improving performance of Jaccard coefficient by compensating for its drawbacks. We conducted experiments using datasets of various characteristics for performance analysis. As a result of comparison between the proposed and the similarity metric of Pearson correlation widely used up to date, it is found that the two metrics yielded competitive performance on a dense dataset while the proposed showed much better performance on a sparser dataset. Also, the result of comparing the proposed with Jaccard coefficient showed that the proposed yielded far better performance as the dataset is denser. Overall, the proposed coefficient demonstrated the best prediction and recommendation performance among the experimented metrics.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권2호
• /
• pp.832-854
• /
• 2019

Action recognition has been studied in computer vision field for years. We present an effective approach to recognize actions using a dimension reduction method, which is applied as a crucial step to reduce the dimensionality of feature descriptors after extracting features. We propose to use sparse matrix and randomized kd-tree to modify it and then propose modified Local Fisher Discriminant Analysis (mLFDA) method which greatly reduces the required memory and accelerate the standard Local Fisher Discriminant Analysis. For feature encoding, we propose a useful encoding method called mix encoding which combines Fisher vector encoding and locality-constrained linear coding to get the final video representations. In order to add more meaningful features to the process of action recognition, the convolutional neural network is utilized and combined with mix encoding to produce the deep network feature. Experimental results show that our algorithm is a competitive method on KTH dataset, HMDB51 dataset and UCF101 dataset when combining all these methods.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2023년도 추계학술발표대회
• /
• pp.124-125
• /
• 2023

Defect detection during liquid crystal display (LCD) manufacturing has always been a critical challenge. This study aims to address this issue by proposing a data augmentation method based on generative adversarial networks (GAN) to improve defect identification accuracy in LCD production. By leveraging synthetically generated image data from GAN, we effectively augment the original dataset to make it more representative and diverse. This data augmentation strategy enhances the model's generalization capability and robustness on real-world data. Compared to traditional data augmentation techniques, the synthetic data from GAN are more realistic, diverse and broadly distributed. Experimental results demonstrate that training models with GAN-generated data combined with the original dataset significantly improves the detection accuracy of critical defects in LCD manufacturing, compared to using the original dataset alone. This study provides an effective data augmentation approach for intelligent quality control in LCD production.