In this study, we performed algorithms to predict algae of Chlorophyll-a (Chl-a). Water quality and quantity data of the middle Nakdong River area were used. At first, the correlation analysis between Chl-a and water quality and quantity data was studied. We extracted ten factors of high importance for water quality and quantity data about the two weirs. Algorithms predicted how ten factors affected Chl-a occurrence. We performed algorithms about decision tree, random forest, elastic net, gradient boosting with Python. The root mean square error (RMSE) value was used to evaluate excellent algorithms. The gradient boosting showed 10.55 of RMSE value for the Gangjeonggoryeong (GG) site and 11.43 of RMSE value for the Dalsung (DS) site. The gradient boosting algorithm showed excellent results for GG and DS sites. Prediction value for the four algorithms was also evaluated through the Receiver operating characteristic (ROC) curve and Area under curve (AUC). As a result of the evaluation, the AUC value was 0.877 at GG site and the AUC value was 0.951 at DS site. So the algorithm‘s ability to interpret seemed to be excellent.

PURPOSES : This study analyzes the characteristics of generated fine particulate matter (PM2.5) and nitrogen oxide (NOX) at roadsides using a statistical method, namely, a generalized linear model (GLM). The study also investigates the applicability and capability of a machine learning methods such as a generalized regression neural network (GRNN) for predicting PM2.5 and NOX generations. METHODS : To analyze the characteristics of PM2.5 and NOX generations at roadsides, data acquisition was conducted in a specific segment of roads, and PM2.5 and NOX prediction models were estimated using GLM. In addition, to investigate the applicability and capability of a machine learning methods, PM2.5 and NOX prediction models were estimated using a GRNN and were compared with models employing previously estimated GLMs using r-square, mean absolute deviation (MAD), mean absolute percentage error (MAPE), and root mean square error (RMSE) as parameters. RESULTS : Results revealed that relative humidity, wind speed, and traffic volume were significant for both PM2.5 and NOX prediction models based on estimated models from a GLM. In addition, to compare the applicability and capability of the GLM and GRNN models (i.e., PM2.5 and NOX prediction models), the GRNN model of PM2.5 and NOX prediction was found to yield better statistical significance for r-square, MAD, MAPE, and RMSE as compared with the same parameters used in the GLM. CONCLUSIONS : Analytical results indicated that a higher relative humidity and traffic volume could lead to higher PM2.5 and NOX concentrations. By contrast, lower wind speed could affect higher PM2.5 and NOX concentrations at roadsides. In addition, based on a comparison of two statistical methods (i.e., GLM and GRNN models used to estimate PM2.5 and NOX), GRNN model yielded better statistical significance as compared with GLM.

Anomaly detection of Machine Learning such as PCA anomaly detection and CNN image classification has been focused on cross-sectional data. In this paper, two approaches has been suggested to apply ML techniques for identifying the failure time of big time series data. PCA anomaly detection to identify time rows as normal or abnormal was suggested by converting subjects identification problem to time domain. CNN image classification was suggested to identify the failure time by re-structuring of time series data, which computed the correlation matrix of one minute data and converted to tiff image format. Also, LASSO, one of feature selection methods, was applied to select the most affecting variables which could identify the failure status. For the empirical study, time series data was collected in seconds from a power generator of 214 components for 25 minutes including 20 minutes before the failure time. The failure time was predicted and detected 9 minutes 17 seconds before the failure time by PCA anomaly detection, but was not detected by the combination of LASSO and PCA because the target variable was binary variable which was assigned on the base of the failure time. CNN image classification with the train data of 10 normal status image and 5 failure status images detected just one minute before.

선박이 접안할 때 발생하는 접안에너지에 가장 영향력이 큰 요소는 접안속도이며, 과도한 경우 사고로 이어질 수 있다. 접안속도의 결정에 영향을 미치는 요소는 다양하지만 기존 연구에서는 일반적으로 선박 크기에 제한하여 분석하였다. 따라서 본 연구에서는 다양한 선박 접안속도의 영향요소를 반영하여 분석하고 그에 따른 중요도를 도출하고자 한다. 분석에 활용한 데이터는 국내 한 탱커부두의 선박 접안속도를 실측한 것을 바탕으로 하였다. 수집된 데이터를 활용하여 머신러닝 분류 알고리즘인 의사결정나무(Decision Tree), 랜덤포 레스트(Random Forest), 로지스틱회귀(Logistic Regression), 퍼셉트론(Perceptron)을 비교분석하였다. 알고리즘 평가 방법으로는 혼동 행렬에 따른 모델성능 평가지표를 사용하였다. 분석 결과, 가장 성능이 좋은 알고리즘으로는 퍼셉트론이 채택되었으며 그에 따른 접안속도 영향 요인의 중요도는 선박 크기(DWT), 부두 위치(Jetty No.), 재화상태(State) 순으로 나타났다. 이에 따라 선박 접안 시, 선박의 크기를 비롯하여 부두 위치, 재화 상태 등 다양한 요인을 고려하여 접안속도를 설계하여야 한다.

본 연구는 운항선의 운항 빅데이터를 활용하여 머신러닝 기법의 선박 마력 예측에 관한 것이다. 현재 신조선에는 ISO15016법을 이용하여 외부환경 요인에 대하여 수식을 통해 저항을 예측하나 관련 계산식이 복잡하고 요구하는 입력변수들이 많아 운항하는 실선 적 용에 많은 시간과 비용이 필요하다. 본 연구에서는 최근 예측, 인식 등에서 우수한 성능을 보이는 SVM(Support Vector Machine) 알고리즘을 이용하여 우수한 성능의 선박 출력 예측이 가능한 모델을 제안한다. 제안 예측 모델은 실선 운항 빅데이터만 확보된다면 ISO15016법 대비 우수한 성능의 예측이 가능한 장점이 있다. 본 연구에서는 178K 벌크캐리어의 운항 DATA를 활용하여 ISO15016 기법과 본 연구에서 제안 하는 SVM 알고리즘 기반의 마력해석법을 비교하여 ISO15016의 단점인 선박 모델 데이터 준비 부분을 줄이고 부정확한 마력 예측 성능을 개선하였다.

인공지능은 우리가 매일 들고 다니는 휴대폰을 비롯한 생활 전반뿐만 아니라 모든 사회 영역에까지 침투하고 있다. 특히 최근에 대두된 인공지능의 학습방법으로서의 머신러닝 기법의 발달로 이제 인공지능은 엄청난 양의 데이터를 학습하면서 성장해나가는 형태를 기본으로 하고 있다. 사법절차를 비롯한 분쟁해결영역도 인공지능이 넘보는 분야에서 더 이상 예외가 아니다. 언뜻 생각하면 법체계는 언어, 논리, 그리고 개념 간의 관련성이 지배하는 영역이어서, 전자적⋅기계적인 분석방법이 적극적으로 도입될 수 있는 분야로 비추어지나, 실제로 아직까지는 법 영역에서 자연어처리나 머신러닝 기법은 기대만큼의 성과를 내지는 못하고 있다. 최근 판결의 결과를 예측하는 시스템을 개발하는 연구들이 활발한데, 현재까지 공표된 대부분의 판결예측 연구들은 일단 사실관계는 확정을 해둔 채, 해당 사실관계의 특징적인 요소들이 담긴 기존 사례들을 분석하여 인간이 하는 것과 유사한 결론을 도출하는 과정으로 이루어진다. 그에 비하여 날 것 그대로의 증거들을 분석하여 허위의 주장이나 증거들을 걸러내고 진실한 사실관계를 재구성하는 정도에 이른 연구들은 찾아보기 힘들다. 특히 우리나라의 경우에는 인공지능이 학습에 필요한 충분한 판례 데이터를 확보하지 못하기가 어렵다는 점이 많이 지적되고 있는 현실이다. 다만 판결문을 훈련데이터로 활용하게 되더라도, 이 유기재가 생략된 판결문도 많다는 점이나 판사 개인이 판결 이유를 기재하는 방식이 훈련데이터의 분석에 영향을 끼칠 수도 있다는 점 등을 유념 해야 한다. 또한 판결문 등 기존 훈련데이터에 편향성이나 오류가 내재되어 있다면, 인공지능의 학습도 당연히 영향을 받을 수밖에 없다. 나아가 인공지능이 추상적인 불확정개념의 해석이나 적용 까지도 할 수 있을지는 아직 의문이다. 인공지능 판사의 출현은 적어도 단시일에 이루어지지는 않을 것이고, 단기적으로는 인공지능을 활용하여 판사의 업무를 보조하는 역할을 더 기대해볼 수 있겠다. 인공지능이 사법분야 내에서 활동할 수 있는 영역을 구체적ㆍ세부적으로 연구 하는 작업은, 인간 법관이 주도권을 잡은 채로 인공지능이 보조적인 역할을 하는 시대에서 인공지능을 활용할 수 있는 영역을 효과적으로 짚어내는 데에 도움이 될 수 있다. 다만 점차적으로 판사를 보조하는 지위에서 어느새 조금씩 인간 판사의 판단을 잠식할 위험성은 항시 경계해야 한다. 특히 인공지능 시스템이 개선될수록 판사가 인공지능이 작성한 보고서나 판결문 초안에 과도하게 의지하여 기계적, 통계적인 판단만을 내리게 될 위험을 조심해야 한다.

개(Canis lupus familiaris)는 인간의 소외 현상을 개선하고, 공동체 생활 의식 향상에 기여하는 반려동물이다. 반려견 품종을 명확히 관리하는 것은 유전병을 감소시키거나, 형질 개량, 종 다양성 유지 등을 위해 중요하다. 본 연구에서는 고밀도 SNP 칩 유전자형 데이터와 기계학습 기술을 이용하여, 유전자형 데이터에 기반한 품종 식별이 가능한지, 가능하다면 최소 몇 개의 유전마커로 품종 식별을 유의하게 수행할 수 있는지 확인하기 위하여, 반려견 11 품종 226두의 23K SNP 칩 데이터를 분석하였다. 9종의 기계학습 다중범주 분류 알고리즘과 2종의 특징 선택 방법의 성능을 비교하여, 선형 서포트 벡터 머신 분류기와 주성분 분석 특징 기여도를 이용한 특징 선택 방법을 이용했을 때, 11종의 반려견 품종을 90% 이상 정확도로 식별하였으며, 이 때 40개의 유전마커가 필요함을 확인하였다. 최종 선발 된 40개의 반려견 품종 식별 유전마커는 타 질병 예측 마커와 결합하여 유전자 검사 키트로 제작될 수 있으며, 반려견 품종 관리 및 질병 관리 기술로 유용하게 활용될 수 있을 것이다.

In this research, we evaluate on the disassemblability of recycling process for vehicle front door using the symbolic chart method and machine-learning algorithm. It is applied to the front door of 1600cc class vehicle, and then the conventional steel door and CFRP door were compared. Based on the principle symbolic chart method, the number of processes can be different according to decomposer proficiency of suitability of recycling process, so the evaluation method is required to supply this issue. The machine learning algorithm, and artificial intelligence method were applied and the applicable tools for each experiment were used to compensate the variations in the number of processes according to different proficiencies. Because CFRP front door has integrated components compare to steel door, so its disassemblability processes were decreased to 80 from 103 of the conventional steel door’s. It can be confirmed that the disassemblability was increased from the suitability of recycling equation. In case of the steel, disassemblability was approximately 60.6, in case of the CFRP is approximately 72 for car front door. Therefore, it can be concluded that the disassemblability of CFRP was better in the evaluation of suitability of recycling.

It is known that the growth and development of the mosquito are greatly affected by the change of the meteorological factors. In particular, temperature and precipitation are closely related to the life cycle of the mosquito, and their effects have different characteristics for each species of mosquito. Therefore, to develop a mosquito activity index based on mosquito density, it is essential to develop a prediction model based on weather data. In this study, we developed a functional formula that can estimate the change of mosquito density according to the change of meteorological factors using the mosquito classification data of Incheon region collected from 2011 to 2017. Also, using the data of the digital mosquito monitoring system (DMS) from April to October 2018, mosquito activity index according to characteristics of space in city was developed. In order to reflect the temporal characteristics of the mosquito life-cycle, we used data of temperature and precipitation prior to 1-2 weeks, and used land cover data to reflect the spatial characteristics of mosquito density. Density of Culex pipiens collected in the Incheon area were gradually increased when the average temperature increased two weeks ago after adjusting the precipitation. However, when the average temperature reached 22°C, the density of Culex pipiens marked a peak, and above the 22°C, the density was decreased. The predicted mosquito activity index calculated by applying the machine learning method to the DMS data of the Incheon area is designed to calculate from 1 to 10 grades. The accuracy of the mosquito activity index was 87% when the one grade error was allowed.

With the advent of the digital age, production and distribution of web pages has been exploding. Internet users frequently need to extract specific information they want from these vast web pages. However, it takes lots of time and effort for users to find a specific information in many web pages. While search engines that are commonly used provide users with web pages containing the information they are looking for on the Internet, additional time and efforts are required to find the specific information among extensive search results. Therefore, it is necessary to develop algorithms that can automatically extract specific information in web pages. Every year, thousands of international conference are held all over the world. Each international conference has a website and provides general information for the conference such as the date of the event, the venue, greeting, the abstract submission deadline for a paper, the date of the registration, etc. It is not easy for researchers to catch the abstract submission deadline quickly because it is displayed in various formats from conference to conference and frequently updated. This study focuses on the issue of extracting abstract submission deadlines from International conference websites. In this study, we use three machine learning models such as SVM, decision trees, and artificial neural network to develop algorithms to extract an abstract submission deadline in an international conference website. Performances of the suggested algorithms are evaluated using 2,200 conference websites.

PURPOSES : The purpose of this study is to compare applicability, explanation power, and flexibility of traffic accident models between estimating model using the statistical method and the machine learning method. METHODS: In order to compare and analyze traffic accident models between model estimated using the statistical method and machine learning method, data acquisition was conducted, and traffic accident models were estimated using statistical methods such as negative binomial regression model, and machine learning methods such as a generalized regression neural network (GRNN). Then, the fitness of model as R2, root mean square error (RMSE), mean absolute percentage error (MAPE), accuracy, etc., were determined to compare the traffic accident models. RESULTS: The results showed that the annual average daily traffic (AADT), speed limits, number of lanes, land usage, exclusive right turn lanes, and front signals were significant for both traffic accident models. The GRNN model of total traffic accidents had been better statistical significant with R2: 0.829, RMSE: 2.495, MAPE: 32.158, and Accuracy: 66.761 compared with the negative binomial regression model with R2: 0.363, RMSE: 9.033, MAPE: 68.987, and Accuracy: 8.807. The GRNN model of injury traffic accidents also showed similar results of model’s statistical significance. CONCLUSIONS: Traffic accident models estimated with GRNN had better statistical significance compared with models estimated with statistical methods such as negative binomial regression model.