• 경영정보학연구
• /
• 제22권4호
• /
• pp.21-39
• /
• 2020

딥러닝(Deep learning) 기법은 패턴분석, 이미지분류 등 다양한 분야에서 높은 성과를 나타내고 있다. 특히, 주식시장 분석문제는 머신러닝 연구분야에서도 어려운 분야이므로 딥러닝이 많이 활용되는 영역이다. 본 연구에서는 패턴분석과 분류능력이 높은 딥러닝의 일종인 합성곱신경망(Convolutional Neural Network) 모델을 활용하여 주가방향 예측방법을 제안한다. 추가적으로 합성곱신경망 모델을 효율적으로 학습시키기 위한 속성선택(Feature Selection, FS)방법이 적용된다. 합성곱신경망 모델의 성과는 머신러닝 단일 분류기와 앙상블 분류기를 벤치마킹하여 객관적으로 검증된다. 본 연구에서 벤치마킹한 분류기는 로지스틱 회귀분석(Logistic Regression), 의사결정나무(Decision Tree), 인공신경망(Neural Network), 서포트 벡터머신(Support Vector Machine), 아다부스트(Adaboost), 배깅(Bagging), 랜덤포레스트(Random Forest)이다. 실증분석 결과, 속성선택을 적용한 합성곱신경망이 다른 벤치마킹 분류기보다 분류 성능이 상대적으로 높게 나타났다. 이러한 결과는 합성곱신경망 모델과 속성선택방법을 적용한 예측방법이 기업의 재무자료에 내포된 가치를 보다 정교하게 분석할 수 있는 가능성이 있음을 실증적으로 확인할 수 있었다.

• 대한치과기공학회지
• /
• 제7권1호
• /
• pp.5-12
• /
• 1985

This experimental study is perform to study the effect of the direction for the centrifugal force in operating centrifugal casting machine. In order to investigate the effect of direction for centrifugal force action on the casting ability 20 pieces of specimen are produced including one vertical direction and four horizontal directions. Casting shows us that the 15 pieces of specimen among them are included within $\pm$16. The results obtained through the comparison with 5 groups of the casting ability are as follows: 1. In the experiment of the casting, statistically significant difference are found between the vertical direction and the group of horizontal direction, namely, the direction of 3, 6, 9 and 12 o'clock, and also statistically there are significant differences between the direction of 9 o'clock and of 6 o'clock, and between the direction of 9 o'clock and of 12 o'clock which lie between the groups of horizontal directions. 2. The degree of casting ability is showing us on an average, 23.80 in the vertical direction. And also in the group of horizontal direction it prove in turn 15.60 in the direction of 9 o'clock, 14.47 of 3 o'clock, 13.30 of 12 o'clock, and 12.80 of 6 o'clock.

• 한국기계가공학회지
• /
• 제13권1호
• /
• pp.52-58
• /
• 2014

This study consists of two parts. The first discusses the development of a single production forming machine which was reported in earlier papers. The second outlines the development of a multi-production forming machine, which consists primarily of a film feeding unit, an unwinding unit, and a heating block unit. The heating block unit of the multi-production forming machine has 30 members per die. An analysis of the stress deformation and temperature deviation of this machine is carried out using ANSYS Workbench and CFX-11 under the design conditions. According to this analysis, the maximum deflection in the Z-direction is $0.05104{\mu}m$ and the maximum temperature deviation is $0.7^{\circ}C$ when the temperature of the heating block unit is $175^{\circ}C$. It was also found that these values are structurally safe. The advantage of the developed multi-production forming machine is demonstrated to be in its offering of a proper voice test.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2002년도 춘계학술대회 논문집
• /
• pp.632-635
• /
• 2002

In this paper, a general procedure to determine machine tool errors from the on-machine measurement (OMM) data is described. First, a parameterized error model of a machine tool is illustrated by approximating error components as linear function of axis positions, and a modified error model is proposed which includes backlash effects. To determine the unknown model coefficient vectors of the forward and backward error model, an artifact with 8 cutes is made and calibrated on CMM. Then, lower-left and upper-right cube corners are measured with a touch-trigger probe mounted on the machine tool spindle. Measured error data are used to determine the coefficient vectors. The positioning errors in the XY plane at the fixed z position are simulated for the forward and backward error model.

• 한국공작기계학회:학술대회논문집
• /
• 한국공작기계학회 2001년도 춘계학술대회 논문집(한국공작기계학회)
• /
• pp.46-52
• /
• 2001

One of the important technical issues is how to decrease thermal expansion of ballscrew in proportion to the increase of machining speed. when measuring force of stretch of ballscrew, since not only actual expansion and the value of bending have to be considered, it's impossible to definite the exact value of expansion. In addition, support bearings of ballscrew gain considerable force in axial direction. It also generates thermal expansion on the ballscrew, and deteriorates the bearings. In conclusion, it's impossible to give the pretension enough to absorb the all elongation due to thermal expansion generated during machine running. If gave, bed, column and saddle are all bent to change machine accuracy, and the support bearings of ballscrew are damaged. The purpose of this paper is to study the pretension of support bearing of ballscrew in machine tool.

• 한국공작기계학회:학술대회논문집
• /
• 한국공작기계학회 1999년도 추계학술대회 논문집 - 한국공작기계학회
• /
• pp.440-445
• /
• 1999

One of the important technical issues is how to decrease thermal expansion of ballscrew in proportion to the increase of machining speed. When measuring force of stretch of ballscrew, since not only actual expansion and the value of bending have to be considered, it's impossible to definite the exact value of expansion. In addition, support bearings of ballscrew gain considerable force in axial direction. It also generates thermal expansion on the ballscrew, and deteriorates the bearings. In conclusion, it's impossible to give the pretension enough to absorb the all elongation due to thermal expansion generated during machine running. If gave, bed, column and saddle are all bent to change machine accuracy, and the pretension of support bearing of ballscrew in machine tool.

• 한국공작기계학회논문집
• /
• 제11권3호
• /
• pp.51-56
• /
• 2002

In order to achieve high precision machine tools, Performance enhancement of feed drive systems is required. One of the important technical issues is how to decrease thermal expansion of ball screw in proportion to the increase of machining speed. When measuring force of stretch of ball screw, since not only actual expansion and the value of bending have to be considered, it is impossible to define the exact value of expansion. In addition, support bearings of ball screw gain considerable force in axial direction. It also generates thermal expansion on the ball screw, and deteriorates the performances of the hearings. In conclusion, it is impossible to give the pretension enough to absorb all the elongation due to thermal expansion generated during machine is running. If given bed column and saddle are all bent to chance machine accuracy, and the support bearings of ball screw is damaged.

• 한국생산제조학회지
• /
• 제19권6호
• /
• pp.746-752
• /
• 2010

Rib structure has been used to bed and column of machine tool to heighten weight stiffness ratio, cost performance and weight saving. In this paper the bed rib structure was estimated with thermal characteristics. Using superposition principle, machine tool designer can describe every complicated heat generation in the machine tool thermal source. As thermal characteristics, thermal deformation of guide rail and column and Maximum-minimum temperature variation were selected. The size, configuration and direction against the thermal loading surface operated to the thermal characteristics. The DB chart was made following rib structure estimating thermal characteristics. With superposition principle and DB chart, machine tool designer can prognosticate the thermal characteristics without FEM analysis to every bed rib structure.

• 서비스연구
• /
• 제13권3호
• /
• pp.105-126
• /
• 2023

국내 중고차 시장은 지속적으로 성장하고 있으며, 이와 동시에 중고차 온라인 플랫폼 서비스 역시 함께 매년 시장 점유율을 확대하고 있다. 중고차 온라인 플랫폼 서비스는 차량의 제원, 점검 이력, 사고 내역, 그리고 세부 옵션 등을 서비스 이용자에게 제공하고 있다. 대부분의 기존 연구는 차량의 제원과 차량의 일부 옵션을 활용한 중고차 가격의 예측이었으며, 중고차 가격과 일부 제원 변수 간 비선형 관계임을 확인하였다. 이에 따라 연구자들은 이러한 비선형 문제를 해결하기 위해 머신러닝(Machine Learning) 모델의 실행을 제안하였으며, 그 결과 회귀(Regression) 기반 머신러닝 모델은 변수의 실질적인 영향력과 방향성을 알 수 있는 장점이 존재하였으나, 트리(Decision Tree) 기반 머신러닝 모델에 비해 비용함수 수치가 저조한 단점이 존재하였다. 본 연구는 국내 브랜드를 대상으로 차량의 제원과 차량의 옵션, 총 70여 개의 변수를 모두 활용하여 회귀 기반 머신러닝 모델과 트리 기반 머신러닝 모델을 순차적으로 실행하여 두 유형의 머신러닝 모델의 장점을 취합하고자 하였다. 이를 통해 브랜드별 변수의 실질적 영향력과 방향성을 확인한 후 브랜드별 가장 우수한 트리 기반 머신러닝 모델을 선정하였다. 본 연구의 시사점은 다음과 같다. 중고차 온라인 플랫폼 서비스를 이용하는 구매자와 판매자가 전반적인 중고차 가격 예측을 지원할 수 있다. 이에 따라 중고차 온라인 플랫폼 서비스 이용자 간 정보의 비대칭으로 인한 문제 해결 역시 지원이 가능할 것으로 기대한다.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2000년도 춘계학술대회 논문집
• /
• pp.676-680
• /
• 2000

In this paper, an extended volumetric error model considering backlash in a three-axis machine tool was proposed and utilized for calculating the volumetric error of the machine tool at any position in three-dimensional workspace. Backlashes are interrelated; i.e. the angular backlash affects the straightness errors which then affect the calculated squareness errors. Therefore, a new concept was introduced to define the backlash of squareness errors to incorporate the backlash of squareness error into the volumetric error, and the characteristics of the backlash of squareness error were investigated. The effects of backlash errors were assessed, by experiments, fur 21 geometric errors of a machine tool. The backlash error was shown to be one of the systematic errors of a machine tool. Based on this volumetric error model, a computer-aided volumetric error analysis system was developed for a three-axis machine tool in this paper. Then the volumetric error at an arbitrary position can be obtained, and displayed in a three-dimensional graphic form.