• The KIPS Transactions:PartB
• /
• v.10B no.6
• /
• pp.719-724
• /
• 2003

This paper proposed a heuristic to enhance the speed of kNN without loss of its accuracy. The proposed heuristic minimizes the computation of the similarity between two documents which is the dominant factor in kNN. To do this, the paper proposes a method to calculate the upper limit of the similarity and to sort the training documents. The proposed heuristic was implemented on the existing framework of the text categorization, so called, AI :: Categorizer and it was compared with the conventional kNN with the well-known data, Router-21578. The comparisons show that the proposed heuristic outperforms kNN about 30∼40% with respect to the execution time.

• Animal Bioscience
• /
• v.37 no.4
• /
• pp.622-630
• /
• 2024

Objective: Pig breeders cannot obtain phenotypic information at the time of selection for sow lifetime productivity (SLP). They would benefit from obtaining genetic information of candidate sows. Genomic data interpreted using deep learning (DL) techniques could contribute to the genetic improvement of SLP to maximize farm profitability because DL models capture nonlinear genetic effects such as dominance and epistasis more efficiently than conventional genomic prediction methods based on linear models. This study aimed to investigate the usefulness of DL for the genomic prediction of two SLP-related traits; lifetime number of litters (LNL) and lifetime pig production (LPP). Methods: Two bivariate DL models, convolutional neural network (CNN) and local convolutional neural network (LCNN), were compared with conventional bivariate linear models (i.e., genomic best linear unbiased prediction, Bayesian ridge regression, Bayes A, and Bayes B). Phenotype and pedigree data were collected from 40,011 sows that had husbandry records. Among these, 3,652 pigs were genotyped using the PorcineSNP60K BeadChip. Results: The best predictive correlation for LNL was obtained with CNN (0.28), followed by LCNN (0.26) and conventional linear models (approximately 0.21). For LPP, the best predictive correlation was also obtained with CNN (0.29), followed by LCNN (0.27) and conventional linear models (approximately 0.25). A similar trend was observed with the mean squared error of prediction for the SLP traits. Conclusion: This study provides an example of a CNN that can outperform against the linear model-based genomic prediction approaches when the nonlinear interaction components are important because LNL and LPP exhibited strong epistatic interaction components. Additionally, our results suggest that applying bivariate DL models could also contribute to the prediction accuracy by utilizing the genetic correlation between LNL and LPP.

• Animal Bioscience
• /
• v.37 no.5
• /
• pp.826-831
• /
• 2024

Objective: The major histocompatibility complex in chicken demonstrates a great range of variations within varities, breeds, populations and that can eventually influence their immuneresponses. The preset study was conducted to understand the major histocompatibility complex-B (MHC-B) variability in five major populations of Bangladesh native chicken: Aseel, Hilly, Junglefowl, Non-descript Deshi, and Naked Neck. Methods: These five major populations of Bangladesh native chicken were analyzed with a subset of 89 single nucleotide polymorphisms (SNPs) in the high-density MHC-B SNP panel and Kompetitive Allele-Specific polymerase chain reaction genotyping was applied. To explore haplotype diversity within these populations, the results were analyzed both manually and computationally using PHASE 2.1 program. The phylogenetic investigations were also performed using MrBayes program. Results: A total of 136 unique haplotypes were identified within these five Bangladesh chicken populations, and only one was shared (between Hilly and Naked Neck). Phylogenetic analysis showed no distinct haplotype clustering among the five populations, although they were shared in distinct clades; notably, the first clade lacked Naked Neck haplotypes. Conclusion: The present study discovered a set of unique MHC-B haplotypes in Bangladesh chickens that could possibly cause varied immune reponses. However, further investigations are required to evaluate their relationships with global chicken populations.

• Communications for Statistical Applications and Methods
• /
• v.9 no.3
• /
• pp.889-898
• /
• 2002

In decision theoretic estimation, the loss function usually emphasizes precision of estimation. However, one may have interest in goodness of fit of the overall model as well as precision of estimation. From this viewpoint, Zellner(1994) proposed the balanced loss function which takes account of both "goodness of fit" and "precision of estimation". This paper considers estimation of the parameter of a Bernoulli distribution using Zellner's(1994) balanced loss function. It is shown that the sample mean $\overline{X}$, is admissible. More general results, concerning the admissibility of estimators of the form $a\overline{X}+b$ are also presented. Finally, minimax estimators and some numerical results are given at the end of paper,at the end of paper.

• Proceedings of the Korean Information Science Society Conference
• /
• 2006.10c
• /
• pp.240-245
• /
• 2006

본 논문에서는 의상에 대한 사용자 선호도를 찾아내는 기법에 대하여 기술한다. 의상에 대한 사용자 선호도를 찾기 위해서 의상 데이터에 대해 데이터 모델을 새롭게 제안한다. 이 데이터 모델을 기반으로 사용자의 의상관련 히스토리를 저장한다. 이렇게 저장된 히스토리 정보에 기계 학습 기법 중 최근 각광받고 있는 SVM 기법을 적용하여 사용자 선호도를 찾아내도록 하였다. 이 결과를 다른 학습 기법인 Naive Bayes 기법을 사용하여 의상에 대한 사용자 선호도를 검색한 성능과 비교하여 우리 모델이 더 좋다는 것을 확인하였다. 우리는 5명의 사용자에 대해서 동일한 취향을 갖는 사용자가 몇 명인지에 따라 A(모두 다름), B(2명), C(3명), D(4명), E(모두 같음) 형태별, 사용자별 1000건의 히스토리를 일정한 기준에 따라 생성했다. 그리고 이 중에서 900건을 학습용 데이터, 100건을 검증용 데이터로 선정하여 실험이 진행되었다.

• Communications for Statistical Applications and Methods
• /
• v.24 no.2
• /
• pp.115-127
• /
• 2017

This paper proposes a Bayesian test for the equivalence of survival functions in multiple groups. Proposed Bayesian test use the model of Cox's regression with time-varying coefficients. B-spline expansions are used for the time-varying coefficients, and the proposed test use only the partial likelihood, which provides easier computations. Various simulations of the proposed test and typical tests such as log-rank and Fleming and Harrington tests were conducted. This result shows that the proposed test is consistent as data size increase. Specifically, the power of the proposed test is high despite the existence of crossing hazards. The proposed test is based on a Bayesian approach, which is more flexible when used in multiple tests. The proposed test can therefore perform various tests simultaneously. Real data analysis of Larynx Cancer Data was conducted to assess applicability.

• Proceedings of the KOSOMBE Conference
• /
• v.1995 no.05
• /
• pp.49-53
• /
• 1995

In this paper, we classify between the chromosome and blood cell, and find the location of chromosome. First, the gray level images be the binary images using the threshold method. Then, the spot noises are removed by the morphological filtering. Features are obtained using the updated Run length(RL) coding and are classified using the Bayes decision rule. The performances of classification are 83.3% in chromosome and 93.3% in blood cell. Because each sub-images ($256{\times}256$) is obtained from the full image($512{\times}512$), we realize the location of chromosome if we get the corrected chromosome classifications.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.05a
• /
• pp.501-504
• /
• 2021

한국어 심층학습 언어모델인 KoBERT와, 확률적 기계학습 분류기인 나이브 베이즈와 로지스틱 회귀를 이용하여 유학생이 작성한 한국어 쓰기 답안지의 점수 구간을 예측하는 실험을 진행하였다. 네가지 주제('직업', '행복', '경제', '성공')를 다룬 답안지와 점수 레이블(A, B, C, D)로 쌍을 이룬 학습데이터 총 304건으로 다양한 자동분류 모델을 구축하여 7-겹 교차검증을 시행한 결과 KoBERT가 나이브 베이즈나 로지스틱 회귀보다 약간 우세한 성능을 보였다.

• ETRI Journal
• /
• v.45 no.6
• /
• pp.1065-1078
• /
• 2023

In vehicular networks, diverse safety information can be shared among vehicles through internet connections. In vehicle-to-internet communications, vehicles on the road are wirelessly connected to different cloud networks, thereby accelerating safety information exchange. Onboard sensors acquire traffic-related information, and reliable intermediate nodes and network services, such as navigational facilities, allow to transmit safety information to distant target vehicles and stations. Using vehicle-to-network communications, we minimize delays and achieve high accuracy through consistent connectivity links. Our proposed approach uses intermediate nodes with two-hop separation to forward information. Target vehicle detection and routing of safety information are performed using machine learning algorithms. Compared with existing vehicle-to-internet solutions, our approach provides substantial improvements by reducing latency, packet drop, and overhead.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.31B no.7
• /
• pp.170-178
• /
• 1994

The nearest neighbor classification rule is widely used because it is not only simple but the error rate is asymptotically less than twice Bayes theoretical minimum error. But the method basically use the whole training patterns as the reference vectors. so that both storage and classification time increase as the number of training patterns increases. LVQ(Learning Vector Quantization) resolved this problem by training the reference vectors instead of just storing the whole training patterns. But it is a heuristic algorithm which has no theoretic background there is no terminating condition and it requires a lot of iterations to get to meaningful result. This paper is to propose a new training method of the reference vectors. which minimize the given error function. The nearest neighbor network,the network version of the nearest neighbor classification rule is proposed. The network is funtionally identical to the nearest neighbor classification rule is proposed. The network is funtionally identical to the nearest neighbor classification rule and the reference vectors are represented by the weights between the nodes. The network is trained to minimize the error function with respect to the weights by the steepest descent method. The learning algorithm is derived and it is shown that the proposed method can adjust more reference vectors than LVQ in each iteration. Experiment showed that the proposed method requires less iterations and the error rate is smaller than that of LVQ2.