• Journal of Animal Science and Technology
• /
• v.62 no.2
• /
• pp.247-262
• /
• 2020

Birth weight and subsequent weight gain is of critical importance in the survival and performance of piglets on a commercial swine farm setting. Oropharyngeal microbiome could influence immunity, and feeding behavior thus impacting health and weight gain. We used 16S rRNA gene sequencing to profile the composition and predicted metabolic functionality of the oropharyngeal microbiota in 8 piglets (4 with a birthweight ≤ 1.0 kg and 4 with a birthweight ≥ 1.7 kg) at 11, 26, and 63 days of age. We found 9 genera that were significantly associated with average daily gain (ADG) at 11 days (false discovery rate, FDR < 0.05) and 26 days of age (FDR < 0.1), respectively. The microbial functional profile revealed several pathways associated with ADG (FDR < 0.05). Among these, pathways related to degradation of catechols showed a positive association with ADG at 11, 26, and 63 days of age, implying a potential to breakdown the host-derived catecholamines. We also noted that pathways related to the biodegradation of nucleosides and nucleotides increased with ADG during the pre-weaning phase, while those involved in their biosynthesis decreased. Our findings provide insights into the oropharyngeal microbial memberships and metabolic pathways that are involved in a piglet's weight gain. Thus, providing a basis for the development of strategies aimed at improving weight gain in pigs.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.2
• /
• pp.959-981
• /
• 2017

Existing Android malware detection approaches mostly have concentrated on superficial features such as requested or used permissions, which can't reflect the essential differences between benign apps and malware. In this paper, we propose a quantitative calculation model of application risks based on the key observation that the essential differences between benign apps and malware actually lie in the way how permissions are used, or rather the way how their corresponding permission methods are used. Specifically, we employ a fine-grained analysis on Android application risks. We firstly classify application risks into five specific categories and then introduce comprehensive risk, which is computed based on the former five, to describe the overall risk of an application. Given that users' risk preference and risk-bearing ability are naturally fuzzy, we design and implement a fuzzy logic system to calculate the comprehensive risk. On the basis of the quantitative calculation model, we propose a risk classification based approach for Android malware detection. The experiments show that our approach can achieve high accuracy with a low false positive rate using the RandomForest algorithm.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.14 no.1
• /
• pp.35-43
• /
• 2018

MANET composed of only wireless nodes is increasingly utilized in various fields. However, it is exposed to many security vulnerabilities because it doesn't have any infrastructure and transmits data by using multi-hop method. Therefore, MANET should be applied the intrusion detection technique that can detect efficiently malicious nodes and decrease impacts of various attacks. In this paper, we propose a distributed intrusion detection technique that can detect the various attacks while improving the efficiency of attack detection and reducing the false positive rate. The proposed technique uses the cluster structure to manage the information in the center and monitor the traffic of their neighbor nodes directly in all nodes. We use three parameters for attack detection. We also applied an efficient authentication technique using only key exchange without the help of CA in order to provide integrity when exchanging information between cluster heads. This makes it possible to free the forgery of information about trust information of the nodes and attack nodes. The superiority of the proposed technique can be confirmed through comparative experiments with existing intrusion detection techniques.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.9
• /
• pp.4271-4294
• /
• 2018

Communication cost is the most important factor in Wireless Sensor Networks (WSNs), as exchanging control keying messages consumes a large amount of energy from the constituent sensor nodes. Time-based Dynamic Keying and En-Route Filtering (TICK) can reduce the communication costs by utilizing local time values of the en-route nodes to generate one-time dynamic keys that are used to encrypt reports in a manner that further avoids the regular keying or re-keying of messages. Although TICK is more energy efficient, it employs no re-encryption operation strategy that cannot determine whether a healthy report might be considered as malicious if the clock drift between the source node and the forwarding node is too large. Secure SOurce-BAsed Loose Synchronization (SOBAS) employs a selective encryption en-route in which fixed nodes are selected to re-encrypt the data. Therefore, the selection of encryption nodes is non-adaptive, and the dynamic network conditions (i.e., The residual energy of en-route nodes, hop count, and false positive rate) are also not focused in SOBAS. We propose an energy efficient selection of re-encryption nodes based on fuzzy logic. Simulation results indicate that the proposed method achieves better energy conservation at the en-route nodes along the path when compared to TICK and SOBAS.

• Annual Conference on Human and Language Technology
• /
• 2004.10d
• /
• pp.214-219
• /
• 2004

전자 메일은 매우 많은 사람들이 사용하는 편리하고 효율적인 통신 수단이다. 그러나 전자메일 주소를 쉽게 획득할 수 있다면 점을 악용하기 때문에 사용자가 원하지 않는 메일 즉 스팸 메일에 대한 문제가 심각해지고 있다. 이러한 스팸 메일을 자동으로 분류해주는 스팸 필터는 주로 영어를 대상으로 하고 있으며, 규칙 기반 필터링보다는 통계적 학습을 통한 필터링 방법을 주로 사용하고 있다. 본 논문에서는 베이즈 정리를 기반으로 하는 3가지 분류 알고리즘을 한글 전자메일을 대상으로 하여 스팸 메일 특히 음란성 메일을 분류하는데 있어 그 성능을 평가하고자 한다. 실험 결과, 단어의 스팸일 확률만을 이용하는 방법이 나이브 베이즈 알고리즘이나 m-estimate를 이용하는 방법보다는 성능이 우수함을 알 수 있었다 특히, 단어의 스팸일 확률만을 이용하는 방법은 false positive rate를 0%로 유지하면서도 다른 방법들보다는 필터링을 잘 해내고 있음을 확인할 수 있었다. 그리고, 자질 선정에서는 명사나 명사/형용사를 사용할 경우에 그 에러율이 가장 적었다.

• Journal of Information Processing Systems
• /
• v.13 no.4
• /
• pp.1014-1028
• /
• 2017

Social networking services (SNSs) such as Twitter, MySpace, and Facebook have become progressively significant with its billions of users. Still, alongside this increase is an increase in security threats such as cross-site scripting (XSS) threat. Recently, a few approaches have been proposed to detect an XSS attack on SNSs. Due to the certain recent features of SNSs webpages such as JavaScript and AJAX, however, the existing approaches are not efficient in combating XSS attack on SNSs. In this paper, we propose a machine learning-based approach to detecting XSS attack on SNSs. In our approach, the detection of XSS attack is performed based on three features: URLs, webpage, and SNSs. A dataset is prepared by collecting 1,000 SNSs webpages and extracting the features from these webpages. Ten different machine learning classifiers are used on a prepared dataset to classify webpages into two categories: XSS or non-XSS. To validate the efficiency of the proposed approach, we evaluated and compared it with other existing approaches. The evaluation results show that our approach attains better performance in the SNS environment, recording the highest accuracy of 0.972 and lowest false positive rate of 0.87.

• Journal of IKEEE
• /
• v.22 no.4
• /
• pp.1175-1179
• /
• 2018

In this paper, we propose recurrent CNN(Convolutional Neural Networks) for detecting seizures among patients using EEG signals. In the proposed method, data were mapped by image to preserve the spectral characteristics of the EEG signal and the position of the electrode. After the spectral preprocessing, we input it into CNN and extracted the spatial and temporal features without wavelet transform. Results from the Children's Hospital of Boston Massachusetts Institute of Technology (CHB-MIT) dataset showed a sensitivity of 90% and a false positive rate (FPR) of 0.85 per hour.

• Toxicological Research
• /
• v.35 no.4
• /
• pp.343-351
• /
• 2019

Many in vitro developmental toxicity assays have been proposed over several decades. Since the late 1980s, we have made intermittent attempts to introduce in vitro assays as screening tests for developmental toxicity of inhouse candidate products. Two cell-based assays which were developed two decades apart were intensively studied. One was an assay of inhibitory effects on mouse ascites tumor cell attachment to a concanavalin A-coated plastic sheet surface (MOT assay), which we studied in the early days of assay development. The other was an assay of inhibitory effects on the differentiation of mouse embryonic stem cell to beating heart cells (EST assay), which we assessed more recently. We evaluated the suitability of the assays for screening in-house candidates. The concordance rates with in vivo developmental toxicity were at the 60% level. The EST assay classified chemicals that inhibited cell proliferation as embryo-toxic. Both assays had a significant false positive rate. The assays were generally considered unsuitable for screening the developmental toxicity of our candidate compounds. Recent test systems adopt advanced technologies. Despite such evolution of materials and methods, the concordance rates of the EST and MOT systems were similar. This may suggest that the fundamental predictivity of in vitro developmental toxicity assays has remained basically unchanged for decades. To improve their predictivity, in vitro developmental toxicity assays should be strictly based on elucidated pathogenetic mechanisms of developmental toxicity.

• Genomics & Informatics
• /
• v.20 no.1
• /
• pp.9.1-9.6
• /
• 2022

Mendelian randomization (MR) uses genetic variation as a natural experiment to investigate the causal effects of modifiable risk factors (exposures) on outcomes. Two-sample Mendelian randomization (2SMR) is widely used to measure causal effects between exposures and outcomes via genome-wide association studies. 2SMR can increase statistical power by utilizing summary statistics from large consortia such as the UK Biobank. However, the first-order term approximation of standard error is commonly used when applying 2SMR. This approximation can underestimate the variance of causal effects in MR, which can lead to an increased false-positive rate. An alternative is to use the second-order approximation of the standard error, which can considerably correct for the deviation of the first-order approximation. In this study, we simulated MR to show the degree to which the first-order approximation underestimates the variance. We show that depending on the specific situation, the first-order approximation can underestimate the variance almost by half when compared to the true variance, whereas the second-order approximation is robust and accurate.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.11
• /
• pp.4105-4121
• /
• 2021

Recently, malware classification based on Deep Neural Networks (DNN) has gained significant attention due to the rise in popularity of artificial intelligence (AI). DNN-based malware classifiers are a novel solution to combat never-before-seen malware families because this approach is able to classify malwares based on structural characteristics rather than requiring particular signatures like traditional malware classifiers. However, these DNN-based classifiers have been found to lack robustness against malwares that are carefully crafted to evade detection. These specially crafted pieces of malware are referred to as adversarial examples. We consider a clever adversary who has a thorough knowledge of DNN-based malware classifiers and will exploit it to generate a crafty malware to fool DNN-based classifiers. In this paper, we propose a DNN-based malware classifier that becomes resilient to these kinds of attacks by exploiting Generative Adversarial Network (GAN) based data augmentation. The experimental results show that the proposed scheme classifies malware, including AEs, with a false positive rate (FPR) of 3.0% and a balanced accuracy of 70.16%. These are respective 26.1% and 18.5% enhancements when compared to a traditional DNN-based classifier that does not exploit GAN.