• Journal of The Korean Society of Agricultural Engineers
• /
• v.59 no.3
• /
• pp.97-110
• /
• 2017

This study investigates the efficiencies of machine learning models, including artificial neural network (ANN), generalized regression neural network (GRNN), adaptive neuro-fuzzy inference system (ANFIS) and random forest (RF), for reservoir water level forecasting in the Chungju Dam, South Korea. The models' efficiencies are assessed based on model efficiency indices and graphical comparison. The forecasting results of the models are dependent on lead times and the combination of input variables. For lead time t = 1 day, ANFIS1 and ANN6 models yield superior forecasting results to RF6 and GRNN6 models. For lead time t = 5 days, ANN1 and RF6 models produce better forecasting results than ANFIS1 and GRNN3 models. For lead time t = 10 days, ANN3 and RF1 models perform better than ANFIS3 and GRNN3 models. It is found that ANN model yields the best performance for all lead times, in terms of model efficiency and graphical comparison. These results indicate that the optimal combination of input variables and forecasting models depending on lead times should be applied in reservoir water level forecasting, instead of the single combination of input variables and forecasting models for all lead times.

• Structural Engineering and Mechanics
• /
• v.77 no.2
• /
• pp.231-243
• /
• 2021

In this paper, a dynamic response mapping model of the wheel-rail system is established by using the support vector regression (SVR) method, and the hierarchical safety thresholds of the subgrade void are proposed based on the reliability theory. Firstly, the vehicle-track coupling dynamic model considering the subgrade void is constructed. Secondly, the subgrade void area, the subgrade compaction index K30 and the fastener stiffness are selected as random variables, and the mapping model between these three random parameters and the dynamic response of the wheel-rail system is built by using the orthogonal test and the SVR. The sensitivity analysis is carried out by the range analysis method. Finally, the hierarchical safety thresholds for the subgrade void are proposed. The results show that the subgrade void has the most significant influence on the carbody vertical acceleration, the rail vertical displacement, the vertical displacement and the slab tensile stress. From the range analysis, the subgrade void area has the largest effect on the dynamic response of the wheel-rail system, followed by the fastener stiffness and the subgrade compaction index K30. The recommended safety thresholds for the subgrade void of level I, II and III are 4.01㎡, 6.81㎡ and 9.79㎡, respectively.

• International Journal of Computer Science & Network Security
• /
• v.23 no.8
• /
• pp.190-198
• /
• 2023

To enhance customer satisfaction for higher profits, an e-commerce sector can establish a continuous relationship and acquire new customers. Utilize machine-learning models to analyse their customer's behavioural evidence to produce their competitive advantage to the e-commerce platform by helping to improve overall satisfaction. These models will forecast customers who will churn and churn causes. Forecasts are used to build unique business strategies and services offers. This work is intended to develop a machine-learning model that can accurately forecast retainable customers of the entire e-commerce customer data. Developing predictive models classifying different imbalanced data effectively is a major challenge in collected data and machine learning algorithms. Build a machine learning model for solving class imbalance and forecast customers. The satisfaction accuracy is used for this research as evaluation metrics. This paper aims to enable to evaluate the use of different machine learning models utilized to forecast satisfaction. For this research paper are selected three analytical methods come from various classifications of learning. Classifier Selection, the efficiency of various classifiers like Random Forest, Logistic Regression, SVM, and Gradient Boosting Algorithm. Models have been used for a dataset of 8000 records of e-commerce websites and apps. Results indicate the best accuracy in determining satisfaction class with both gradient-boosting algorithm classifications. The results showed maximum accuracy compared to other algorithms, including Gradient Boosting Algorithm, Support Vector Machine Algorithm, Random Forest Algorithm, and logistic regression Algorithm. The best model developed for this paper to forecast satisfaction customers and accuracy achieve 88 %.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.156-156
• /
• 2022

Explicitly spatially distributed and reliable data on industrial water demand is very much important for both policy makers and researchers in order to carry a region-specific analysis of water resources management. However, such type of data remains scarce particularly in underdeveloped and developing countries. Current research is limited in using different spatially available socio-economic, climate data and geographical data from different sources in accordance to predict industrial water demand at finer resolution. This study proposes a random forest regression (RFR) model to predict the industrial water demand at 0.50× 0.50 spatial resolution by combining various features extracted from multiple data sources. The dataset used here include National Polar-orbiting Partnership (NPP)/Visible Infrared Imaging Radiometer Suite (VIIRS) night-time light (NTL), Global Power Plant database, AQUASTAT country-wise industrial water use data, Elevation data, Gross Domestic Product (GDP), Road density, Crop land, Population, Precipitation, Temperature, and Aridity. Compared with traditional regression algorithms, RF shows the advantages of high prediction accuracy, not requiring assumptions of a prior probability distribution, and the capacity to analyses variable importance. The final RF model was fitted using the parameter settings of ntree = 300 and mtry = 2. As a result, determinate coefficients value of 0.547 is achieved. The variable importance of the independent variables e.g. night light data, elevation data, GDP and population data used in the training purpose of RF model plays the major role in predicting the industrial water demand.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2009.01a
• /
• pp.223-228
• /
• 2009

We propose an on-line machine learning approach for object recognition, where new images are continuously added and the recognition decision is made without delay. Random forest (RF) classifier has been extensively used as a generative model for classification and regression applications. We extend this technique for the task of building incremental component-based detector. First we employ object descriptor model based on bag of covariance matrices, to represent an object region then run our on-line RF learner to select object descriptors and to learn an object classifier. Experiments of the object recognition are provided to verify the effectiveness of the proposed approach. Results demonstrate that the propose model yields in object recognition performance comparable to the benchmark standard RF, AdaBoost, and SVM classifiers.

• Communications for Statistical Applications and Methods
• /
• v.13 no.1
• /
• pp.177-190
• /
• 2006

The aim of this study is to propose a Bayesian model for fitting mortality rate of colon cancer. For the analysis of mortality rate of a disease, factors such as age classes of population and spatial characteristics of the location are very important. The model proposed in this study allows the age class to be a random effect in addition to its conventional role as the covariate of a linear regression, while the spatial factor being a random effect. The model is fitted using Metropolis-Hastings algorithm. Posterior expected predictive deviances, standardized residuals, and residual plots are used for comparison of models. It is found that the proposed model has smaller residuals and better predictive accuracy. Lastly, we described patterns in disease maps for colon cancer.

• Asian-Australasian Journal of Animal Sciences
• /
• v.29 no.5
• /
• pp.607-614
• /
• 2016

The objectives of the study were to estimate genetic parameters for milk production traits of Holstein cattle using random regression models (RRMs), and to compare the goodness of fit of various RRMs with homogeneous and heterogeneous residual variances. A total of 126,980 test-day milk production records of the first parity Holstein cows between 2007 and 2014 from the Dairy Cattle Improvement Center of National Agricultural Cooperative Federation in South Korea were used. These records included milk yield (MILK), fat yield (FAT), protein yield (PROT), and solids-not-fat yield (SNF). The statistical models included random effects of genetic and permanent environments using Legendre polynomials (LP) of the third to fifth order (L3-L5), fixed effects of herd-test day, year-season at calving, and a fixed regression for the test-day record (third to fifth order). The residual variances in the models were either homogeneous (HOM) or heterogeneous (15 classes, HET15; 60 classes, HET60). A total of nine models (3 orders of $polynomials{\times}3$ types of residual variance) including L3-HOM, L3-HET15, L3-HET60, L4-HOM, L4-HET15, L4-HET60, L5-HOM, L5-HET15, and L5-HET60 were compared using Akaike information criteria (AIC) and/or Schwarz Bayesian information criteria (BIC) statistics to identify the model(s) of best fit for their respective traits. The lowest BIC value was observed for the models L5-HET15 (MILK; PROT; SNF) and L4-HET15 (FAT), which fit the best. In general, the BIC values of HET15 models for a particular polynomial order was lower than that of the HET60 model in most cases. This implies that the orders of LP and types of residual variances affect the goodness of models. Also, the heterogeneity of residual variances should be considered for the test-day analysis. The heritability estimates of from the best fitted models ranged from 0.08 to 0.15 for MILK, 0.06 to 0.14 for FAT, 0.08 to 0.12 for PROT, and 0.07 to 0.13 for SNF according to days in milk of first lactation. Genetic variances for studied traits tended to decrease during the earlier stages of lactation, which were followed by increases in the middle and decreases further at the end of lactation. With regards to the fitness of the models and the differential genetic parameters across the lactation stages, we could estimate genetic parameters more accurately from RRMs than from lactation models. Therefore, we suggest using RRMs in place of lactation models to make national dairy cattle genetic evaluations for milk production traits in Korea.

• Journal of Animal Science and Technology
• /
• v.47 no.2
• /
• pp.147-158
• /
• 2005

The purpose of this study was to find out the effects that daughters' incomplete lactation records affect sire's breeding values through genetic evaluation using RRTDM(random regression test-day model). First, we estimated genetic parameters and breeding values on sires having complete lactation records of daughter by RRTDM, second, we changed complete lactation records of specific sires into incomplete records by various methods. Third, the breeding values were compared between complete and incomplete records. Finally, this study aimed to find out the methods to minimize the estimation errors of young bulls' breeding values. Data used in this study were collected from the dairy herd improvement program, and a total of 97,562 records were composed of 10,929 first parity with both parents known, since 1999. Breeding values on the daughters from randomly chosen sires were calculated and compared with among 90 day, 150day, and 200 day's incomplete records. For milk yields, sire's ranks of breeding values used by complete lactation records were very different from sire's ranks of breeding values obtained by incomplete lactation records(Rank_90 cut, 150cut, 200 cut).The differences were also obtained between complete lactation records(per305_full) and incomplete lactation record (per_90 cut, 150cut, 200 cut) in breeding values regarding persistency. Especially, the differences between per_90 cut and per305_full were very large(from 1.8 kg to 145kg).

• Communications for Statistical Applications and Methods
• /
• v.29 no.2
• /
• pp.239-250
• /
• 2022

In many applications, we frequently encounter correlated multiple outcomes measured on the same subject. Joint modeling of such multiple outcomes can improve efficiency of inference compared to independent modeling. For instance, in developmental toxicity studies, fetal weight and number of malformed pups are measured on the pregnant dams exposed to different levels of a toxic substance, in which the association between such outcomes should be taken into account in the model. The number of malformations may possibly have many zeros, which should be analyzed via zero-inflated count models. Motivated by applications in developmental toxicity studies, we propose a Bayesian joint modeling framework for continuous and count outcomes with excess zeros. In our model, zero-inflated Poisson (ZIP) regression model would be used to describe count data, and a subject-specific random effects would account for the correlation across the two outcomes. We implement a Bayesian approach using MCMC procedure with data augmentation method and adaptive rejection sampling. We apply our proposed model to dose-response analysis in a developmental toxicity study to estimate the benchmark dose in a risk assessment.

• Journal of Korea Water Resources Association
• /
• v.55 no.spc1
• /
• pp.1283-1293
• /
• 2022

The purpose of this study is to predict residual chlorine in order to maintain stable residual chlorine concentration in sedimentation basin by using artificial intelligence algorithms in water treatment process employing pre-chlorination. Available water quantity and quality data are collected and analyzed statistically to apply into mathematical multiple regression and artificial intelligence models including multi-layer perceptron neural network, random forest, long short term memory (LSTM) algorithms. Water temperature, turbidity, pH, conductivity, flow rate, alkalinity and pre-chlorination dosage data are used as the input parameters to develop prediction models. As results, it is presented that the random forest algorithm shows the most moderate prediction result among four cases, which are long short term memory, multi-layer perceptron, multiple regression including random forest. Especially, it is result that the multiple regression model can not represent the residual chlorine with the input parameters which varies independently with seasonal change, numerical scale and dimension difference between quantity and quality. For this reason, random forest model is more appropriate for predict water qualities than other algorithms, which is classified into decision tree type algorithm. Also, it is expected that real time prediction by artificial intelligence models can play role of the stable operation of residual chlorine in water treatment plant including pre-chlorination process.