• Journal of Distribution Research
• /
• v.12 no.3
• /
• pp.69-90
• /
• 2007

This study is set out to investigate the factors that influence customers' behavior of choice and switching between online and offline channels, separating the purchase decision into two stages, i.e., information search and purchase. Factors influencing channel choice are found to differ from stage to stage. The main results of this study are as follows. At the information search stage, customers' channel knowledge had impacts on the choice of the channel. Customers are more likely to visit offline bookstores when they have hedonic shopping orientation and higher involvement level with books. On the contrary, customers are more apt to search online when they have a lot of online shopping experiences. At the purchase stage, the results varied according to the search channel. When customers search for information online, the following variables lead to online purchases: online shopping experiences with books, price-focused shopping orientation, and time availability for shopping. Perceived risk made customers purchase offline even though they searched online. In case of offline searching, customers with more convenience-focused, hedonic-focused shopping orientation and less tim availability purchased offline.

• Journal of Biosystems Engineering
• /
• v.38 no.2
• /
• pp.95-102
• /
• 2013

Purpose: Using a laboratory indented cylinder separator, broken rice separation experiments were conducted and the characteristics of the separation process were studied to provide information for developing a prototype indented cylinder broken rice separator. Methods: Rice (Ilmi variety) milled in a local RPC was used for the experiment. Rice kernels were classified into four groups according to their length l; whole kernels (I > 3.75 mm), semi-whole kernels (2.5 < I < 3.75 mm), broken kernels (1.75 < I < 2.5 mm), and foreign matters (I < 1.75 mm). A laboratory grain cleaner, Labofix '90 (Schmidt AG, Germany) was used for the experiments. Experiments were designed as a $4{\times}4$ factorial arrangement in randomized blocks with three replications. Cylinder rotational speeds (17, 34, 51, 68 rpm) and trough angles (15, 37.5, 60, $82.5^{\circ}$) were the two factors and feed rates (25, 50 kg/h), indent shapes (Us, $S_1$ type), and indent sizes (2.5, 3.75 mm) were treated as the blocks. Two 125 g samples and one 125 g sample were taken at the cylinder outlet and from the trough, respectively. The whole, semi-whole, and broken kernel weight ratio of the samples and feed was determined by a rice sizing device. From these weight ratios, purities, degrees of extraction and coefficient of separation efficiency were calculated. Results: Trough angle, cylinder speed, and their interaction on the coefficient of separation efficiency were statistically significant. Cylinder speed of 17, 34, and 51 rpm made the most effective separation when the trough angle was $15^{\circ}$ or $37.5^{\circ}$, $60^{\circ}$, and $82.5^{\circ}$, respectively. Maximum values of coefficient of separation efficiency were in the range of 60 to 70% except when the indent size was 2.5 mm and were recorded for the combinations of low cylinder speed (17 rpm) with medium trough angle ($37.5^{\circ}$ or $60^{\circ}$). Indent shape did not appear to make any noticeable difference in separation efficiency. Conclusions: Due to the interaction effect, the trough angle needs to be increased appropriately when an increase in cylinder speed is made if a rapid drop of effectiveness of separation should be avoided. In commercial applications, $S_1$ type indents are preferred because of their better manufacturability and easier maintenance. For successful separation of broken kernels, the indent size should be set slightly bigger than the actual sizes of broken kernels: an indent size of 3.0 mm for separating broken kernels shorter than 2.5 mm.

• Journal of Intelligence and Information Systems
• /
• v.18 no.2
• /
• pp.29-45
• /
• 2012

Bond rating is regarded as an important event for measuring financial risk of companies and for determining the investment returns of investors. As a result, it has been a popular research topic for researchers to predict companies' credit ratings by applying statistical and machine learning techniques. The statistical techniques, including multiple regression, multiple discriminant analysis (MDA), logistic models (LOGIT), and probit analysis, have been traditionally used in bond rating. However, one major drawback is that it should be based on strict assumptions. Such strict assumptions include linearity, normality, independence among predictor variables and pre-existing functional forms relating the criterion variablesand the predictor variables. Those strict assumptions of traditional statistics have limited their application to the real world. Machine learning techniques also used in bond rating prediction models include decision trees (DT), neural networks (NN), and Support Vector Machine (SVM). Especially, SVM is recognized as a new and promising classification and regression analysis method. SVM learns a separating hyperplane that can maximize the margin between two categories. SVM is simple enough to be analyzed mathematical, and leads to high performance in practical applications. SVM implements the structuralrisk minimization principle and searches to minimize an upper bound of the generalization error. In addition, the solution of SVM may be a global optimum and thus, overfitting is unlikely to occur with SVM. In addition, SVM does not require too many data sample for training since it builds prediction models by only using some representative sample near the boundaries called support vectors. A number of experimental researches have indicated that SVM has been successfully applied in a variety of pattern recognition fields. However, there are three major drawbacks that can be potential causes for degrading SVM's performance. First, SVM is originally proposed for solving binary-class classification problems. Methods for combining SVMs for multi-class classification such as One-Against-One, One-Against-All have been proposed, but they do not improve the performance in multi-class classification problem as much as SVM for binary-class classification. Second, approximation algorithms (e.g. decomposition methods, sequential minimal optimization algorithm) could be used for effective multi-class computation to reduce computation time, but it could deteriorate classification performance. Third, the difficulty in multi-class prediction problems is in data imbalance problem that can occur when the number of instances in one class greatly outnumbers the number of instances in the other class. Such data sets often cause a default classifier to be built due to skewed boundary and thus the reduction in the classification accuracy of such a classifier. SVM ensemble learning is one of machine learning methods to cope with the above drawbacks. Ensemble learning is a method for improving the performance of classification and prediction algorithms. AdaBoost is one of the widely used ensemble learning techniques. It constructs a composite classifier by sequentially training classifiers while increasing weight on the misclassified observations through iterations. The observations that are incorrectly predicted by previous classifiers are chosen more often than examples that are correctly predicted. Thus Boosting attempts to produce new classifiers that are better able to predict examples for which the current ensemble's performance is poor. In this way, it can reinforce the training of the misclassified observations of the minority class. This paper proposes a multiclass Geometric Mean-based Boosting (MGM-Boost) to resolve multiclass prediction problem. Since MGM-Boost introduces the notion of geometric mean into AdaBoost, it can perform learning process considering the geometric mean-based accuracy and errors of multiclass. This study applies MGM-Boost to the real-world bond rating case for Korean companies to examine the feasibility of MGM-Boost. 10-fold cross validations for threetimes with different random seeds are performed in order to ensure that the comparison among three different classifiers does not happen by chance. For each of 10-fold cross validation, the entire data set is first partitioned into tenequal-sized sets, and then each set is in turn used as the test set while the classifier trains on the other nine sets. That is, cross-validated folds have been tested independently of each algorithm. Through these steps, we have obtained the results for classifiers on each of the 30 experiments. In the comparison of arithmetic mean-based prediction accuracy between individual classifiers, MGM-Boost (52.95%) shows higher prediction accuracy than both AdaBoost (51.69%) and SVM (49.47%). MGM-Boost (28.12%) also shows the higher prediction accuracy than AdaBoost (24.65%) and SVM (15.42%)in terms of geometric mean-based prediction accuracy. T-test is used to examine whether the performance of each classifiers for 30 folds is significantly different. The results indicate that performance of MGM-Boost is significantly different from AdaBoost and SVM classifiers at 1% level. These results mean that MGM-Boost can provide robust and stable solutions to multi-classproblems such as bond rating.

• Korean Journal of Environment and Ecology
• /
• v.25 no.2
• /
• pp.211-226
• /
• 2011

The purpose of this paper was to ensure the basis for effective management of Tricholoma matsutake mountain province, to perform biotope mapping of Pinus densiflora based on growth environment of Tricholoma matsutake, target a cluster of Yangyang-gun, Kang Won-do. Study Methods were to review on growth and environmental characteristics of Tricholoma matsutake through internal and external documents and to identify vegetational structure and soil characteristics. This paper studied growth structure and soil environment of Pinus densiflora forest where a farm of production area for Tricholoma matsutake of in order to set the standard of Pinus densiflora biotope. Mapping standards were derived by separating of landform conditions, soil conditions, vegetation conditions. Biotope types were divided into possible production area for Tricholoma matsutake and potential production area for Tricholoma matsutake, possible production area for Tricholoma matsutake were Pinus densiflora biotope in landform and soil structure that enables Tricholoma matsutake production and Single-layered Pinus densiflora biotope of less than 30cm(DBH)-Tree species that other shrub is dominant in shrub layer, Multi-layered Pinus densiflora biotope that Pinus densiflora forest was predominant in understrory layer. Potential production area for Tricholoma matsutake were single-layered Pinus densiflora biotope of more than 30cm(DBH) in landform that enables Tricholoma matsutake production, Pinus densiflora biotope with Quercus predominant in the understrory layer, single-layered Pinus densiflora biotope with Quercus predominant in shrub layer, inappropriate vegetation structure area that the induction of production of Tricholoma matsutake was possible through future vegetation management. According to the research results, Pinus densiflora forest were divided into 16 types; 6 types of possible Tricholoma matsutake production areas, 9 potential Tricholoma matsutake production areas and 16 types of areas where Tricholoma matsutake production was impossible. Possible production areas account for 15.48%, or $9.8km^2$ out of the total Pinus densiflora forest while potential production areas take up 32.42%, or $20.52km^2$, and areas where Tricholoma matsutake production was impossible was 52.10%, or $32.97km^2$.

• Tunnel and Underground Space
• /
• v.28 no.5
• /
• pp.400-425
• /
• 2018

This study presents the research results and current status of the DECOVALEX-2019 project Task B. Task B named 'Fault slip modelling' is aiming at developing a numerical method to simulate the coupled hydro-mechanical behavior of fault, including slip or reactivation, induced by water injection. The first research step of Task B is a benchmark simulation which is designed for the modelling teams to familiarize themselves with the problem and to set up their own codes to reproduce the hydro-mechanical coupling between the fault hydraulic transmissivity and the mechanically-induced displacement. We reproduced the coupled hydro-mechanical process of fault slip using TOUGH-FLAC simulator. The fluid flow along a fault was modelled with solid elements and governed by Darcy's law with the cubic law in TOUGH2, whereas the mechanical behavior of a single fault was represented by creating interface elements between two separating rock blocks in FLAC3D. A methodology to formulate the hydro-mechanical coupling relations of two different hydraulic aperture models and link the solid element of TOUGH2 and the interface element of FLAC3D was suggested. In addition, we developed a coupling module to update the changes in geometric features (mesh) and hydrological properties of fault caused by water injection at every calculation step for TOUGH-FLAC simulator. Then, the transient responses of the fault, including elastic deformation, reactivation, progressive evolutions of pathway, pressure distribution and water injection rate, to stepwise pressurization were examined during the simulations. The results of the simulations suggest that the developed model can provide a reasonable prediction of the hydro-mechanical behavior related to fault reactivation. The numerical model will be enhanced by continuing collaboration and interaction with other research teams of DECOLVAEX-2019 Task B and validated using the field data from fault activation experiments in a further study.

• Journal of The Korean Society of Grassland and Forage Science
• /
• v.36 no.4
• /
• pp.344-349
• /
• 2016

This study was conducted to evaluate the possibility of expanding the usage of whole crop silage from beef cattle and dairy cow to hogs and chickens. For this purpose, a crushing device was developed to crush whole crop silage. The crushed silage was sealed, and analyzed for its feed value. The silage varieties used for the experiment included Saessal barley and Geumgang wheat. Whole crop barley and wheat were crushed in the crushing system as a whole without separating stems, leaves, grains, etc.. When the crushed whole crop silages (CWCS) were analyzed, full grain, grains above 10 mm in size, grains 5~10 mm in size, and grains below 5 mm in size accounted for, 20%, 4%, 27%, and 49 %, respectively. In order to facilitate the fermentation of CWCS, inoculated some fermenter into each CWCS sample (barley or wheat). As control, another set of sample was not inoculated. Crude protein (CP), ether extract (EE), crude fiber (CF), neutral detergent fiber (NDF), acid detergent fiber (ADF), lignin, cellulose content, total digestible nutrient (TDN), and relative feed value (RFV) of fermenter-inoculated Saessal barley were 2.45 %, 1.61%, 8.95%, 16.94%, 9.52%, 1.01%, 8.51%, 81.38%, and 447.5%, respectively. The CP, EE, CF, NDF, ADF, lignin, cellulose content, TDN, and RFV in the other sample of Saessal barley without inoculation of fermenter were 2.57%, 1.62%, 9.61%, 18.25%, 10.13%, 1.10%, 9.04%, 80.90%, and 412.9%, respectively. The CP, EE, CF, NDF, ADF, lignin, cellulose content, TDN, and RFV of fermenter-inoculated Geumgang wheat sample were 2.43%, 1.27%, 10.99%, 19.49%, 11.23%, 1.46%, 9.77%, 80.03%, and 382.6%, respectively. The CP, EE, CF, NDF, ADF, lignin, cellulose content, TDN, RFV of the other set sample of Geumgang wheat sample without the inoculation of fermenter were 2.28%, 1.44%, 10.08%, 18.02%, 10.44%, 1.26%, 9.18%, 80.65%, and 416.9%, respectively. The TDN and RFV content in the fermenter-inoculated Saessal barley were 81.38% and 447.5%, respectively, while the one in the fermenter-inoculated Geumgang wheat were 80.03% and 382.6% respectively. When the feed value of whole crop barley and wheat silage without crushing process was compared to the feed value of whole crop barley and wheat silage made from crushing system, the latter appeared to be higher than the former. This could be due to the process of sealing the crushed silage which might have minimized air content between samples and shortened the golden period of fermentation. In conclusion, these results indicate that a crushing process might be needed to facilitate fermentation and improve the quality of silage when making whole crop silage.

• Journal of Intelligence and Information Systems
• /
• v.16 no.4
• /
• pp.213-225
• /
• 2010

Data mining has empowered the managers who are charge of the tasks in their company to present personalized and differentiated marketing programs to their customers with the rapid growth of information technology. Most studies on customer' response have focused on predicting whether they would respond or not for their marketing promotion as marketing managers have been eager to identify who would respond to their marketing promotion. So many studies utilizing data mining have tried to resolve the binary decision problems such as bankruptcy prediction, network intrusion detection, and fraud detection in credit card usages. The prediction of customer's response has been studied with similar methods mentioned above because the prediction of customer's response is a kind of dichotomous decision problem. In addition, a number of competitive data mining techniques such as neural networks, SVM(support vector machine), decision trees, logit, and genetic algorithms have been applied to the prediction of customer's response for marketing promotion. The marketing managers also have tried to classify their customers with quantitative measures such as recency, frequency, and monetary acquired from their transaction database. The measures mean that their customers came to purchase in recent or old days, how frequent in a period, and how much they spent once. Using segmented customers we proposed an approach that could enable to differentiate customers in the same rating among the segmented customers. Our approach employed support vector regression to forecast the purchase amount of customers for each customer rating. Our study used the sample that included 41,924 customers extracted from DMEF04 Data Set, who purchased at least once in the last two years. We classified customers from first rating to fifth rating based on the purchase amount after giving a marketing promotion. Here, we divided customers into first rating who has a large amount of purchase and fifth rating who are non-respondents for the promotion. Our proposed model forecasted the purchase amount of the customers in the same rating and the marketing managers could make a differentiated and personalized marketing program for each customer even though they were belong to the same rating. In addition, we proposed more efficient learning method by separating the learning samples. We employed two learning methods to compare the performance of proposed learning method with general learning method for SVRs. LMW (Learning Method using Whole data for purchasing customers) is a general learning method for forecasting the purchase amount of customers. And we proposed a method, LMS (Learning Method using Separated data for classification purchasing customers), that makes four different SVR models for each class of customers. To evaluate the performance of models, we calculated MAE (Mean Absolute Error) and MAPE (Mean Absolute Percent Error) for each model to predict the purchase amount of customers. In LMW, the overall performance was 0.670 MAPE and the best performance showed 0.327 MAPE. Generally, the performances of the proposed LMS model were analyzed as more superior compared to the performance of the LMW model. In LMS, we found that the best performance was 0.275 MAPE. The performance of LMS was higher than LMW in each class of customers. After comparing the performance of our proposed method LMS to LMW, our proposed model had more significant performance for forecasting the purchase amount of customers in each class. In addition, our approach will be useful for marketing managers when they need to customers for their promotion. Even if customers were belonging to same class, marketing managers could offer customers a differentiated and personalized marketing promotion.