• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2019.05a
• /
• pp.451-453
• /
• 2019

Recently, as the use of wireless communication devices has increased, the wireless network usage has increased, and a wired network and a wireless network have been mixed to form a network. Existing TCP algorithms are designed for wired networks. Therefore, in the modern network environment, packet loss can not be accurately distinguished and improper congestion control is performed, resulting in degradation of TCP performance. In this paper, we propose SLDA (Support Vector Machine based Loss Discrimination Algorithm) which can accurately classify the packet loss environment to improve TCP performance and evaluate its performance.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.8
• /
• pp.2571-2586
• /
• 2022

The rapid popularity of government social media has generated huge amounts of text data, and the analysis of these data has gradually become the focus of digital government research. This study uses Python language to analyze the big data of the Chinese provincial government Weibo. First, this study uses a web crawler approach to collect and statistically describe over 360,000 data from 31 provincial government microblogs in China, covering the period from January 2018 to April 2022. Second, a word separation engine is constructed and these text data are analyzed using word cloud word frequencies as well as semantic relationships. Finally, the text data were analyzed for sentiment using natural language processing methods, and the text topics were studied using LDA algorithm. The results of this study show that, first, the number and scale of posts on the Chinese government Weibo have grown rapidly. Second, government Weibo has certain social attributes, and the epidemics, people's livelihood, and services have become the focus of government Weibo. Third, the contents of government Weibo account for more than 30% of negative sentiments. The classified topics show that the epidemics and epidemic prevention and control overshadowed the other topics, which inhibits the diversification of government Weibo.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2005.11a
• /
• pp.299-301
• /
• 2005

본 연구는 생체정보를 이용하여 개인을 인증하고 확인하기 위한 방법으로 기존 단일 생체인식 기법의 단점을 보완하기 위해 홍채와 얼굴을 이용한 다중생체인식(Multi-Modal Biometrics Recognition)기법을 연구하였다. 중국 홍채 데이터베이스 CASIA(Chinese Academy of Science)에 Gabor Wavelet과 FLDA(Fuzzy Linear Discriminant Analysis)를 사용하여 특징벡터를 획득하였으며, FERET(FERET(Face Recognition Technology) 얼굴영상데이터를 사용하여 FERET 연구에서 매우 우수한 성능을 보인 EBGM알고리듬으로 특징벡터를 획득하였다. 이로부터 얻어진 두 score 값에 대하여 다양한 균등화 과정을 시도해 보았으며, 등록자와 침입자를 구분하기 위한 Fusion Algorithm으로 Bayesian Classifier, Support vector machine, Fisher's linear discriminant를 사용하였다. 또한, 널리 사용되는 방법 중 Weighted Summation을 이용하여 다중생체인식의 성능을 비교해 보았다.

• MALSORI
• /
• no.66
• /
• pp.61-72
• /
• 2008

This paper proposes a method to improve pathological and normal voice classification performance by combining multiple features such as auditory-based and higher-order features. Their performances are measured by Gaussian mixture models (GMMs) and linear discriminant analysis (LDA). The combination of multiple features proposed by the frame-based LDA method is shown to be an effective method for pathological and normal voice classification, with a 87.0% classification rate. This is a noticeable improvement of 17.72% compared to the MFCC-based GMM algorithm in terms of error reduction.

• Journal of Korea Multimedia Society
• /
• v.23 no.4
• /
• pp.595-602
• /
• 2020

We propose a generative probabilistic model with Dirichlet prior distribution for topic modeling and text similarity analysis. It assigns a topic and calculates text correlation between documents within a corpus. It also provides posterior probabilities that are assigned to each topic of a document based on the prior distribution in the corpus. We then present a Gibbs sampling algorithm for inference about the posterior distribution and compute text correlation among 50 abstracts from the papers published by IEEE. We also conduct a supervised learning to set a benchmark that justifies the performance of the LDA (Latent Dirichlet Allocation). The experiments show that the accuracy for topic assignment to a certain document is 76% for LDA. The results for supervised learning show the accuracy of 61%, the precision of 93% and the f1-score of 96%. A discussion for experimental results indicates a thorough justification based on probabilities, distributions, evaluation metrics and correlation coefficients with respect to topic assignment.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.16 no.3
• /
• pp.25-36
• /
• 2020

Topics calculated using LDA topic modeling have to be labeled separately. When labeling a topic, we look at the words that represent the topic, and label the topic. Therefore, it is important to first make a good set of words that represent the topic. This paper proposes a method of calculating a set of words representing a topic using TextRank, which extracts the keywords of a document. The proposed method uses Relevance to select words related to the topic with discrimination. It extracts topic keywords using the TextRank algorithm and connects keywords with a high frequency of simultaneous occurrence to express the topic with a higher coverage.

• Proceedings of the IEEK Conference
• /
• summer
• /
• pp.435-437
• /
• 2004

This paper presents a personal identification method using teeth images. The method uses images for teeth expressions of anterior and posterior occlusion state and LDA-based technique. Teeth images give merits for recognition because teeth, rigid objects, cannot be deformed at the moment of image acquisition. In the experiments, personal identification for 12 people was successful. It was shown that our method can contribute to multi-modal authentication systems.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.11
• /
• pp.1751-1754
• /
• 2022

The news articles collected with keyword "non face-to-face" were analyzed through topic modeling applied with LDA algorithm. In this study, collected articles were divided into two periods, period 1(the beginning of COVID-19 spread) and period 2(the end of COVID-19 spread), according to issued date of the articles. The articles of period 1 showed support for non-face-to-face treatment, smart library, the beginning of the online financial era, non-face-to-face entrance exam and employment, stock investment for main topic words. And the articles of period 2 showed conversion to non face-to-face classes, increasing unmanned stores, online finance, education industry, home treatment for main topic words. Also, further issues were discussed through visualization of topic words. These results provide evidence that education and unmanned business in non-face-to-face industries are growing.

• Journal of Animal Science and Technology
• /
• v.48 no.3
• /
• pp.337-344
• /
• 2006

The objective of this study was to compare of genetic parameter estimates on carcass traits of Hanwoo(Korean Cattle) according to modeling with Gibbs sampler and MTDFREML. The data set consisted of 1,941 cattle records with 23,058 animals in pedigree files at Hanwoo Improvement Center. The variance and covariance among carcass traits were estimated via Gibbs sampler and MTDFREML algorithms. The carcass traits considered in this study were longissimus dorsi area, backfat thickness, and marbling score. Genetic parameter estimates using Gibbs sampler and MTDFREML from single-trait analysis were similar with those from multiple-trait analysis. The estimated heritabilities using Gibbs sampler were .52～.54, .54 ～.59, and .42～.44 for carcass traits. The estimated heritabilities using MTDFREML were .41, .52～.53, and .31～.32 for carcass traits. The estimated genetic correlation using Gibbs sampler and MTDFREML of LDA between BF and MS were negatively correlated as .34～.36, .23～.37. Otherwise, genetic correlation between BF and MS was positive genetic correlation as .36～.44. The correlations of breeding value for marbling score between via MTDFREML and via Gibbs sampler were 0.989, 0.996 and 0.985 for LDA, BF and MS respectively.

• Progress in Medical Physics
• /
• v.34 no.1
• /
• pp.1-9
• /
• 2023

Purpose: Although ionization chambers are widely used to measure beam commissioning data, point-by-point measurements of all the profiles with various field size and depths are time-consuming tasks. As an alternative, we investigated the feasibility of a linear diode array for commissioning a treatment planning system. Methods: The beam data of a Varian TrueBeam^® radiotherapy system at 6 and 10 MV with/without a flattening filter were measured for commissioning of an Eclipse Analytical Anisotropic Algorithm (AAA) ver.15.6. All of the necessary beam data were measured using an IBA CC13 ionization chamber and validated against Varian "Golden Beam" data. After validation, the measured CC13 profiles were used for commissioning the Eclipse AAA (AAA_CC13). In addition, an IBA LDA-99SC linear diode array detector was used to measure all of the beam profiles and for commissioning a separate model (AAA_LDA99). Finally, the AAA_CC13 and AAA_LDA99 dose calculations for each of the 10 clinical plans were compared. Results: The agreement of the CC13 profiles with the Varian Golden Beam data was confirmed within 1% except in the penumbral region, where ≤2% of a discrepancy related to machine-specific jaw calibration was observed. Since the volume was larger for the CC13 chamber than for the LDA-99SC chamber, the penumbra widths were larger in the CC13 profiles, resulting in ≤5% differences. However, after beam modeling, the penumbral widths agreed within 0.1 mm. Finally the AAA_LDA99 and AAA_CC13 dose distributions agreed within 1% for all voxels inside the body for the 10 clinical plans. Conclusions: In conclusion, the LDA-99SC diode array detector was found to be accurate and efficient for measuring photon beam profiles to commission treatment planning systems.