• The Korean Journal of Applied Statistics
• /
• v.32 no.2
• /
• pp.319-330
• /
• 2019

Forecasting the electricity consumption through analyzing the past electricity consumption a advantageous for energy planing and policy. Machine learning is widely used as a method to predict electricity consumption. Among them, ensemble learning is a method to avoid the overfitting of models and reduce variance to improve prediction accuracy. However, ensemble learning applied to daily data shows the disadvantages of predicting a center value without showing a peak due to the characteristics of ensemble learning. In this study, we overcome the shortcomings of ensemble learning by considering the temperature trend. We compare nine models and propose a model using random forest with the linear trend of temperature.

• Journal of the Korea Society for Simulation
• /
• v.23 no.1
• /
• pp.21-31
• /
• 2014

In this article, data mining simulation is applied to find a proper approach and results of analysis for study of variables related to organization. Also, turnover intention and organizational commitment are used as target (dependent) variables in this simulation. Classification and regression tree (CART) with ensemble methods are used in this study for simulation. Human capital corporate panel data of Korea Research Institute for Vocation Education & Training (KRIVET) is used. The panel data is collected in 2005, 2007, and 2009. Organizational commitment variables are analyzed with combined measure variables which are created after investigation of reliability and single dimensionality for multiple-item measurement details. The results of this study are as follows. First, major determinants of turnover intention are trust, communication, and talent management-oriented trend. Second, the main determining factors for organizational commitment are trust, the number of years worked, innovation, communication. CART with ensemble methods has two ensemble CART methods which are CART with Bagging and CART with Arcing. Comparing two methods, CART with Arcing (Arc-x4) extracted scenarios with very high coefficients of determination. In this study, a scenario with maximum coefficient of determinant and minimum error is obtained and practical implications are presented. Using one of data mining methods, CART with ensemble method. Also, the limitation and future research are discussed.

• Information Systems Review
• /
• v.22 no.4
• /
• pp.21-39
• /
• 2020

Recently, deep learning has shown high performance in various applications such as pattern analysis and image classification. Especially known as a difficult task in the field of machine learning research, stock market forecasting is an area where the effectiveness of deep learning techniques is being verified by many researchers. This study proposed a deep learning Convolutional Neural Network (CNN) model to predict the direction of stock prices. We then used the feature selection method to improve the performance of the model. We compared the performance of machine learning classifiers against CNN. The classifiers used in this study are as follows: Logistic Regression, Decision Tree, Neural Network, Support Vector Machine, Adaboost, Bagging, and Random Forest. The results of this study confirmed that the CNN showed higher performancecompared with other classifiers in the case of feature selection. The results show that the CNN model effectively predicted the stock price direction by analyzing the embedded values of the financial data

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.32 no.3
• /
• pp.17-21
• /
• 2004

Tape-wrapped structures have been generally used in nozzle parts of guided missiles. A continuous band of woven composite material is wrapped around a mandrel that is designed to produce real products. After going through a vacuum bagging process, this woven composite material is cured in a high-pressure autoclave or hydroclave. However, tape-wrapped structures are difficult to analyze because of its large thickness and inclined lay-up. The present study investigates the method of analysis and failure prediction of tape-wrapped structures. The four-point bending test and its finite element analysis were performed to study how to model tape-wrapped structures and investigate their failure characteristics.

• Korean journal of applied entomology
• /
• v.16 no.2 s.31
• /
• pp.127-131
• /
• 1977

This experiment was conducted to evaluate the insect-proof netting, chemical sprays, application of attractants, fruit bagging and light trapping as the control methods of the fruit piercing moths in the orchards on reclaimed land in Sugyeri, Goksung, Chonnam Province, during June to October in 1976. The results are summarized as follows; 1. Insect-proof. netting effectively decreased fruit damage, compared as to the control, down to $9.4\%$ from $38.3\%$ in plum, $2.5\%$ from $53.0\%$ in peaches and $10.0\%$ from $29.0\%$ in grapes. 2. The control effects of chemicals varied significantly among the 7 insecticides tested: Deoclean, Naphthalene, and Thiometon were more effective to the fruit damages as low as $2.0\%,\; 3.6\%,\;and\;5.9\%$ respectively. while the fruit damage was rather high, $9.8\%$ for Demeton, $10.1\%$, for Takju +lead arsenate and $14.2\%$ for Padan. ,3. In the test with 7 attractants, the largest number of moths attracted and killed was 416.by Takju+brown sugar and the next was 307 by Takju+venegor while this number was 141 by mixed solution (see text) which is rather lower than expectation The fruit damage was lowest in Takju+honey and$5.2\%$, the next was $5.60\%$ for Takju+venegor and the highest was $12.0\%$, Takju alone. 4. Fruit bagging with polyethylene film effectively decreased the fruit damage from the inserts but brought about severe fruit rot and delay ripening. Meanwhile, paper bagging was less effective in preventing insects, resulting in $17.5\%$ fruit damage, however, gave no adverse effect other than slight Belay in ripening. 5. Light trapping was hardly expected to be a method of controlling these fruit piercing moths. However, the number of collected moths swarmed by electric light was 10.8 for can-descence, 0.95 for blue, and 0.22 for yellow light.

• Journal of the Korea Society of Computer and Information
• /
• v.21 no.5
• /
• pp.79-89
• /
• 2016

In this paper, we proposed the object classification method using genetic and dynamic random forest consisting of optimal combination of unit tree. The random forest can ensure good generalization performance in combination of large amount of trees by assigning the randomization to the training samples and feature selection, etc. allocated to the decision tree as an ensemble classification model which combines with the unit decision tree based on the bagging. However, the random forest is composed of unit trees randomly, so it can show the excellent classification performance only when the sufficient amounts of trees are combined. There is no quantitative measurement method for the number of trees, and there is no choice but to repeat random tree structure continuously. The proposed algorithm is composed of random forest with a combination of optimal tree while maintaining the generalization performance of random forest. To achieve this, the problem of improving the classification performance was assigned to the optimization problem which found the optimal tree combination. For this end, the genetic algorithm methodology was applied. As a result of experiment, we had found out that the proposed algorithm could improve about 3~5% of classification performance in specific cases like common database and self infrared database compare with the existing random forest. In addition, we had shown that the optimal tree combination was decided at 55~60% level from the maximum trees.

• Korean Journal of Medicinal Crop Science
• /
• v.20 no.5
• /
• pp.365-371
• /
• 2012

This study was carried out to know the characteristics of flowering and bearing fruit, the optimum period, regions and methods for seed harvesting, the optimum temperatures for seed storage and germination, and the optimum period for sowing at nursery bed and seedling transplanting of Valeriana fauriei Briquet. The flowering and bearing fruit of Valeriana fauriei was developed from the before-year root. Optimum period for seed harvest of Valeriana fauriei was from late July to middle August, and optimum areas were the high elevated areas over 500 m above the sea level as Jinbu-myeon, Pyeongchang-gun, Gangwon-do. Using of net-bag for seed harvesting was the effective method to gather the full ripe seed, and bagging of net-bag was necessary from the season of middle May that was the flowering middle-stage. Germination rates don't show the difference among the different temperatures of storage as approximately 41% at $-20^{\circ}C$, $2^{\circ}C$ or $20^{\circ}C$ of seed storage temperatures. The optimum temperature range was in $15{\sim}30^{\circ}C$ for seed germination at nursery bed. The optimum period for seed sowing at nursery bed was the late February, and the optimum period for seedling transplanting was the middle April.

• Journal of the Korea Convergence Society
• /
• v.8 no.10
• /
• pp.45-52
• /
• 2017

Drifter is an equipment for observing the characteristics of seawater in the ocean, and it can be used to predict effluent oil diffusion and to observe ocean currents. In this paper, we design models or the prediction of drifter trajectory using machine learning. We propose methods for estimating the trajectory of drifter using support vector regression, radial basis function network, Gaussian process, multilayer perceptron, and recurrent neural network. When the propose mothods were compared with the existing MOHID numerical model, performance was improve on three of the four cases. In particular, LSTM, the best performed method, showed the imporvement by 47.59% Future work will improve the accuracy by weighting using bagging and boosting.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.11
• /
• pp.41-50
• /
• 2020

Sentimental analysis begins with the search for words that determine the sentimentality inherent in data. Managers can understand market sentimentality by analyzing a number of relevant sentiment words which consumers usually tend to use. In this study, we propose exploring performance of feature selection methods embedded with Particle Swarm Optimization Multi Objectives Evolutionary Algorithms. The performance of the feature selection methods was benchmarked with machine learning classifiers such as Decision Tree, Naive Bayesian Network, Support Vector Machine, Random Forest, Bagging, Random Subspace, and Rotation Forest. Our empirical results of opinion mining revealed that the number of features was significantly reduced and the performance was not hurt. In specific, the Support Vector Machine showed the highest accuracy. Random subspace produced the best AUC results.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.25 no.10
• /
• pp.1514-1519
• /
• 2001

This paper is concerned with design, manufacturing and performance test of LIPCA ( Lightweight Piezo- composite Curved Actuator) using a top carbon fiber composite layer with near -zero CTE(coefficient of thermal expansion), a middle PZT ceramic wafer and a bottom glass/epoxy layer with high CTE. The main point of this design is to replace the heavy metal layers of THUNDER by thigh tweight fiber reinforced plastic layers without losing capabilities to generate high force and large displacement. It is possible to save weight up to about 30% if we replace the metallic backing material by the light fiber composite layer. We can also have design flexibility by selecting the fiber direction and the size of prepreg layers. In addition to the lightweight advantage and design flexibility, the proposed device can be manufactured without adhesive layers when we use epoxy resin prepreg system. Glass/epoxy prepregs, a ceramic wafer with electrode surfaces, and a graphite/epoxy prepreg were simply stacked and cured at an elevated temperature (177 $^{circ}C$ after following an autoclave bagging process. It was found that the manufactured composite laminate device had a sufficient curvature after detached from a flat mold. The analysis method of the cure curvature of LIPCA using the classical lamination theory is presented. The predicted curvatures are fairly in agreement with the experimental ones. In order to investigate the merits of LIPCA, a performance test of both LIPCA and THUNDE$^{TM}$ were conducted under the same boundary conditions. From the experimental actuation tests, it was observed that the developed actuator could generate larger actuation displacement than THUNDERT$^{TM}$.