• Proceedings of the Korea Society of Information Technology Applications Conference
• /
• 2006.06a
• /
• pp.717-729
• /
• 2006

소프트웨어 신뢰도 성장 모델을 비동차 포아송 프로세스(NHPP)로 공식화하고, 이 모델을 이용하여 소프트웨어 신뢰도 척도에 대한 데이터 분석기법을 개발한다. 그간 여러 문헌에서 소프트웨어 신뢰도 성장 모델을 연구할 때 소프트웨어 테스트 중에 소요되는 테스트노력의 양으로서 지수함수형, 레일레이형, 웨이불형, 로지스틱형 테스트노적함수를 고려하여, 경우에 따라 이 중 하나의 적합한 형태를 사용해 왔다. 본 논문에서는 이 네 가지 형태에 대해서 최소자승평가자(LSE)와 최대가능성평가자(MLE)를 써서 파라미터를 구하는 방법을 검토한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.394-395
• /
• 2023

"이성적이며 이상적인 합리적 인간"을 가정하는 기존 경제학의 이론이 항상 실제 상황과 일치하지는 않는 것으로 알려져 있다. 이의 대안으로 나온 행동경제학은, 인간의 경제적 의사결정에 심리, 인지, 감정, 사회문화적 배경 등이 영향을 미친다고 본다. 본 연구에서는 행동경제학에 의거하여, 개인의 감정과 경험이 경제적 의사결정에 영향을 미치는지 여부를 빅데이터 모델을 활용하여 분석하였다. SNS 여론으로는 Reddit, 주가지수로는 S&P 500 을 선정하였다. 수집한 텍스트 데이터를 전처리와 감정분석을 통해 독립변수 값으로 사용했고, 주가지수 등락의 방향성을 종속변수로 사용하여 로지스틱 모형을 구성했다. 모델을 활용하여 분석한 결과 Public sentiment 와 Market sentiment 간 양의 상관관계를 확인할 수 있었다. 또한, lag 를 설정하는 모델이 정확도가 더욱 높음을 확인해, 기존 경제학의 EMH 와 대립되는 바를 확인할 수 있었다. 하지만 최적의 lag 산정을 위해, 더 광범위한 데이터를 바탕으로 한 후속연구가 필요하다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2024.01a
• /
• pp.29-30
• /
• 2024

본 논문은 Krafton의 PUBG: BATTLEGROUNDS 게임에서 플레이어 분류를 목표로 하며, 포즈 추정기술을 사용하여 일반 플레이어와 봇을 구분한다. 이는 게임에서 직접 수집한 비디오 데이터를 기반으로 하며, 다음과 같은 두 가지 접근 방식을 제안한다. 첫 번째 방법은 동작 시퀀스 분석을 통해, 사용자의 특정동작 패턴을 식별하고 로지스틱 회귀 모델을 활용해 사용자 유형을 분류한다. 두 번째 방법은 YOLO-pose 모델을 사용하여 비디오 데이터에서 키포인트를 추출하고, 이를 LSTM 모델에 적용하여 프레임별로 사용자의 유형을 분류한다. 이러한 이중 접근 방식은 게임의 공정성과 사용자 경험을 향상시키는 새로운 도구를 제공하며, 보다 안전한 게임 환경에 기여할 수 있다. 이 연구는 게임 산업뿐만 아니라 보안 및 모니터링 분야에서도 동작 분석에 대한 혁신적인 접근 방식으로 활용될 잠재력을 가지고 있다.

• Advanced Industrial SCIence
• /
• v.3 no.2
• /
• pp.8-16
• /
• 2024

This paper shows the system of drug classification, the goal of this is to foretell the apt drug for the patients based on their demographic and physiological traits. The dataset consists of various attributes like Age, Sex, BP (Blood Pressure), Cholesterol Level, and Na_to_K (Sodium to Potassium ratio), with the objective to determine the kind of drug being given. The models used in this paper are K-Nearest Neighbors (KNN), Logistic Regression and Random Forest. Further to fine-tune hyper parameters using 5-fold cross-validation, GridSearchCV was used and each model was trained and tested on the dataset. To assess the performance of each model both with and without hyper parameter tuning evaluation metrics like accuracy, confusion matrices, and classification reports were used and the accuracy of the models without GridSearchCV was 0.7, 0.875, 0.975 and with GridSearchCV was 0.75, 1.0, 0.975. According to GridSearchCV Logistic Regression is the most suitable model for drug classification among the three-model used followed by the K-Nearest Neighbors. Also, Na_to_K is an essential feature in predicting the outcome.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.13 no.6
• /
• pp.1353-1364
• /
• 2018

This study proposed a method to detect Cochlodinium polykrikoides red tide pixels in satellite images using a logistic regression model of machine learning technique under Imbalanced data. The spectral profiles extracted from red tide, clear water, and turbid water were used as training dataset. 70% of the entire data set was extracted and used for as model training, and the classification accuracy of the model was evaluated using the remaining 30%. At this time, the white noise was added to the spectral profile of the red tide, which has a relatively small number of data compared to the clear water and the turbid water, and over-sampling was performed to solve the unbalanced data problem. As a result of the accuracy evaluation, the proposed algorithm showed about 94% classification accuracy.

• Journal of the Korea Institute of Building Construction
• /
• v.6 no.3 s.21
• /
• pp.99-105
• /
• 2006

In this paper, the estimation of compressive strength of concrete incorporating fly ash subjected to high temperature is discussed. Ordinary Portland cement and fly ash cement(30% of fly ash) were used, respectively. Water to binder ration ranging from 30% to 60% and curing temperature ranging from $20^{\circ}C{\sim}65^{\circ}C$ were also adopted for the experimental parameters. According to results, at the high temperature, FAC had higher strength development at early age than OPC concrete and it kept its high strength development at later age due to accelerated pozzolanic reaction subjected to high temperature. For strength estimation, Logistic model based on maturity equation and Carino model based on equivalent age were applied to verify the availability of estimation model. It shows that fair agreements between calculated values and measured values were obtained evaluating compressive strength with logistic curve. The application of logistic model at high temperature had remarkable deviations in the same maturity. Whereas, the application of Carino model showed good agreements between calculated values and measured ones regardless of type of cement and W/B. However, some correction factors should be considered to enhance the accuracy of strength estimation of concrete.

• Annual Conference on Human and Language Technology
• /
• 2017.10a
• /
• pp.294-297
• /
• 2017

SNS와 스마트기기의 발전으로 온라인을 통한 뉴스 배포가 용이해지면서 악의적으로 조작된 뉴스가 급속도로 생성되어 확산되고 있다. 뉴스 조작은 다양한 형태로 이루어지는데, 이 중에서 정상적인 기사 내에 광고나 낚시성 내용을 포함시켜 독자가 의도하지 않은 정보에 노출되게 하는 형태는 독자가 해당 내용을 진짜 뉴스로 받아들이기 쉽다. 본 논문에서는 뉴스 기사 내에 포함된 문단 중에서 부적합한 문단이 포함되었는지를 판정하기 위한 방법을 제안한다. 제안하는 방식에서는 자연어 처리에 유용한 Convolutional Neural Network(CNN)모델 중 Word2Vec과 tf-idf 알고리즘, 로지스틱 회귀를 함께 이용하여 뉴스 부적합 문단을 검출한다. 본 시스템에서는 로지스틱 회귀를 이용하여 문단의 카테고리를 분류하여 본문의 카테고리 분포도를 계산하고 Word2Vec을 이용하여 문단간의 유사도를 계산한 결과에 가중치를 부여하여 부적합 문단을 검출한다.

• Journal of the Korea Institute of Building Construction
• /
• v.7 no.4
• /
• pp.145-151
• /
• 2007

In this paper, tests are carried out in order to investigate the strength development of concrete under various binder types, W/B and curing temperature ranged from $5{\sim}20^{\circ}C$. Fly ash and blast furnace slag were incorporated by as much as 30%, respectively. Strength development of concrete are estimated using Logistic model and strength ratio of concrete at 28days to that at early age are also investigated. According to experimental results, it is found that good agreements are obtained between measured values and calculated ones using logistic model below $20^{\circ}C$. Strength ratio of concrete at 28days to that at early age increases in case W/B decreases and curing temperature increases. Tables and graphs for strength ratio of concrete are provided in this paper. It is capable of obtaining and predicting the periods to attain design strength by considering increment factor of strength easily with the table and graphs presented in this paper. This paper presents the reference data to decide removal time of form, time to reach target strength and strength inspection of remicon whether the test specimens meet the specified criteria of compressive strength. Multi regression models with respect to the relationship between 7days compressive strength and 28 days compressive strength depending on W/B and admixture types are presented.

• Journal of Advanced Navigation Technology
• /
• v.14 no.1
• /
• pp.119-126
• /
• 2010

The software reliability growth model(SRGM) has been developed in order to estimate such reliability measures as remaining fault number, failure rate and reliability for the developing stage software. Almost of them assumed that the faults detected during testing were eventually removed. Namely, they have studied SRGM based on the assumption that the faults detected during testing were perfectly removed. The fault removing efficiency, however, is imperfect and it is widely known as so in general. It is very difficult to remove detected fault perfectly because the fault detecting is not easy and new error may be introduced during debugging and correcting. Therefore, We want to study imperfect software testing effort for the logistic testing effort which is thought to be the most adequate in this paper.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.20 no.6
• /
• pp.555-561
• /
• 2010

Predicted noise level has been used to assess the annoyance response since noise map was generalized and being the normal method to assess the environmental noise. Unfortunately using predicted noise level to derive the annoyance prediction curve caused some problems. The data have to be grouped manually to use the annoyance prediction curve. The aim of this paper is to propose the method to handle the predicted noise level and the survey data for annoyance prediction curve. This paper used the percentage of persons annoyed(%A) and the percentage of persons highly annoyed as the descriptor of noise annoyance in a population. The logistic regression method was used for deriving annoyance prediction curve. It is concluded that the method of dichotomizing data and logistic regression was suitable to handle the predicted noise level and survey data.