• 지식경영연구
• /
• 제25권2호
• /
• pp.19-47
• /
• 2024

기업의 데이터 분석 및 활용 역량은 전사 차원의 지식경영과 의사결정에 중요한 역할을 한다. 이 연구는 대형 언어 모델이 기업데이터 분석에서 어떻게 활용될 수 있는지 알아보고자 수행되었다. 구체적으로 인적자원 분야에 초점을 맞추어, 대형 언어 모델의 데이터 분석 역량을 검증해 보았다. 이를 위해 인사분야에서 많은 연구가 이루어져온 공개데이터셋 IBM HR 데이터를 소재로, 선행연구들에서 이루어진 머신러닝 기반 이직자 예측 분석을 ChatGPT를 통해 재현하고 그 예측성능을 비교해보았다. 고급 프로그래밍 역량이 필요했던 과거 연구방식과 달리, 분석가의 자연어 요청으로 진행한 ChatGPT 기반 머신러닝 데이터 분석은 훨씬 쉽고 빠르다는 장점이 있었고, 예측 정확도 역시 선행연구와 비교해 경쟁력 있는 수준을 기록했다. 이는 그동안 고급 프로그래밍 역량이 요구되던 기업데이터 분석 분야에서, ChatGPT를 비롯한 대형 언어 모델들이 효과적이고 실질적인 대안이 될 수 있다는 가능성을 시사한다. 또한 이를 통해 데이터 분석의 대중화 나아가 데이터 기반 의사결정(DDDM: Data-Driven Decision Making)의 확산에도 기여할 수 있을 것으로 기대된다. 데이터분석 과정에서 사용한 프롬프트와 ChatGPT가 생성한 프로그래밍 코드도 부록에 수록하여 검증 가능하게 함으로써, 향후 대형 언어 모델을 활용한 데이터분석 연구의 초석을 제공하고자 한다.

• 대한건축학회논문집:구조계
• /
• 제36권3호
• /
• pp.169-176
• /
• 2020

The authors developed a nodal network model that simulates the flow of moist air and the thermal behavior of a target area. The nodal network model was enhanced using a parameter estimation technique based on the measured temperature, humidity, and schedule data. However, the nodal model is not good enough for predicting humidity of the target space, having 55.6% of CVRMSE. It is because re-evaporation effect could not be modeled due to uncertain factors in the field measurement. Hence, a data-driven model was introduced using an artificial neural network (ANN). It was found that the data-driven model is suitable for predicting the condensation compared to the nodal model satisfying ASHRAE Guideline with 3.36% of CVRMSE for temprature, relative humidity, and surface temperature on average. The model will be embedded in automated devices for real-time predictive control, to minimize the risk of surface condensation at dressroom in an apartment housing.

• 한국시뮬레이션학회논문지
• /
• 제5권1호
• /
• pp.1-12
• /
• 1996

We propose a new approach to classifying a time series data into one of the autoregressive moving-average (ARMA) models. It is bases on two pattern recognition concepts for solving time series identification. The one is an extended sample autocorrelation function (ESACF). The other is a neural network-driven decision tree classifier(NNDTC) in which two pattern recognition techniques are tightly coupled : neural network and decision tree classfier. NNDTc consists of a set of nodes at which neural network-driven decision making is made whether the connecting subtrees should be pruned or not. Therefore, time series identification problem can be stated as solving a set of local decisions at nodes. The decision values of the nodes are provided by neural network functions attached to the corresponding nodes. Experimental results with a set of test data and real time series data show that the proposed approach can efficiently identify the time seires patterns with high precision compared to the previous approaches.

• International journal of advanced smart convergence
• /
• 제9권2호
• /
• pp.58-67
• /
• 2020

In this paper, we propose a data-driven-based beam selection scheme for massive multiple-input and multiple-output (MIMO) systems in ultra-dense networks (UDN), which is capable of addressing the problem of high computational cost of conventional coordinated beamforming approaches. We consider highly dense small-cell scenarios with more small cells than mobile stations, in the millimetre-wave band. The analog beam selection for hybrid beamforming is a key issue in realizing millimetre-wave UDN MIMO systems. To reduce the computation complexity for the analog beam selection, in this paper, two deep neural network models are used. The channel samples, channel gains, and radio frequency beamforming vectors between the access points and mobile stations are collected at the central/cloud unit that is connected to all the small-cell access points, and are used to train the networks. The proposed machine-learning-based scheme provides an approach for the effective implementation of massive MIMO system in UDN environment.

• 한국경영과학회지
• /
• 제29권4호
• /
• pp.117-134
• /
• 2004

This study suggests a data driven optimization approach, which simulates the models of human learning processes from cognitive sciences. It shows how the human learning processes can be simulated and applied to solving combinatorial optimization problems. The main advantage of using this method is in applying it into problems, which are very difficult to simulate. 'Undecidable' problems are considered as best possible application areas for this suggested approach. The concept of an 'undecidable' problem is redefined. The learning models in human learning and decision-making related to combinatorial optimization in cognitive and neural sciences are designed, simulated, and implemented to solve an optimization problem. We call this approach 'SLO : simulated learning for optimization.' Two different versions of SLO have been designed: SLO with position & link matrix, and SLO with decomposition algorithm. The methods are tested for traveling salespersons problems to show how these approaches derive new solution empirically. The tests show that simulated learning for optimization produces new solutions with better performance empirically. Its performance, compared to other hill-climbing type methods, is relatively good.

• 한국전산유체공학회지
• /
• 제8권2호
• /
• pp.8-15
• /
• 2003

Three nonlinear κ-ε models with the wall function method are applied to the fully developed turbulent flow in a triangular subchannel of a bare rod bundle. Typical predicted quantities such as axial and secondary velocities, turbulent kinetic energy and wall shear stress are compared in details both qualitatively and quantitatively with both each other and experimental data. The nonlinear κ-ε models by Speziale[1] and Myong and Kasagi[2] are found to be capable of predicting accurately noncircular duct flows involving turbulence-driven secondary motion. The nonlinear κ-ε model by Shih et aL.[3] adopted in a commercial code is found to be unable to predict accurately noncircular flows with the prediction level of secondary flows one order less than that of the experiment.

• 한국구조물진단유지관리공학회지
• /
• 제18권2호
• /
• pp.47-55
• /
• 2014

• Smart Structures and Systems
• /
• 제15권3호
• /
• pp.897-911
• /
• 2015

This study presents a novel approach based on advancements in Evolutionary Computation for data-driven modeling of complex multi-dimensional memory-dependent systems. The investigated example is a benchmark coupled three-dimensional system that incorporates 6 Bouc-Wen elements, and is subjected to external excitations at three points. The proposed technique of this research adapts Genetic Programming for discovering the optimum structure of the differential equation of an auxiliary variable associated with every specific degree-of-freedom of this system that integrates the imposed effect of vibrations at all other degrees-of-freedom. After the termination of the first phase of the optimization process, a system of differential equations is formed that represent the multi-dimensional hysteretic system. Then, the parameters of this system of differential equations are optimized in the second phase using Genetic Algorithms to yield accurate response estimates globally, because the separately obtained differential equations are coupled essentially, and their true performance can be assessed only when the entire system of coupled differential equations is solved. The resultant model after the second phase of optimization is a low-order low-complexity surrogate computational model that represents the investigated three-dimensional memory-dependent system. Hence, this research presents a promising data-driven modeling technique for obtaining optimized representative models for multi-dimensional hysteretic systems that yield reasonably accurate results, and can be generalized to many problems, in various fields, ranging from engineering to economics as well as biology.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 1992년도 수공학연구발표회논문집
• /
• pp.443-447
• /
• 1992

This study describes the strategy for advanced decision support system (ADSS) development for integrated management of water resources and quality in reservoir systems. The developed ADSS consists of database that contain hydrologic data, observed operational data, and data to support specific reservoir operations simulation, optimization models, and water quality models. The optimization model, mass balance simulation model and water quality models are used in a general prototype ADSS, menu driven controlling framework that assists the user to specify and evaluate the alternative operational scenarios at one time. These alternative scenarios are evaluated by the models and the results are compared through the use of a graphical based display system. This graphical based system uses an icon based schematic representation of the system to organize the presentation of the results. The ADSS includes the ability to use monthly or weekly time periods of analysis for the models and it can use monthly historical or stochastically generated inflows.

• International Journal of Aeronautical and Space Sciences
• /
• 제15권4호
• /
• pp.356-365
• /
• 2014

This paper is the second in a series and aims to build a high-fidelity mathematical model for a propeller-driven airplane using the propeller's aerodynamics and inertial models, as developed in the first paper. It focuses on aerodynamic models for the fuselage, the main wing, and the stabilizers under the influence of the wake trailed from the propeller. For this, application of the vortex lattice method is proposed to reflect the propeller's wake effect on those aerodynamic surfaces. By considering the maneuvering flight states and the flow field generated by the propeller wake, the induced velocity at any point on the aerodynamic surfaces can be computed for general flight conditions. Thus, strip theory is well suited to predict the distribution of air loads over wing components and the viscous flow effect can be duly considered using the 2D aerodynamic coefficients for the airfoils used in each wing. These approaches are implemented in building a high-fidelity mathematical model for a propeller-driven airplane. Flight dynamic analysis modules for the trim, linearization, and simulation analyses were developed using the proposed techniques. The flight test results for a series of maneuvering flights with a scaled model were used for comparison with those obtained using the flight dynamics analysis modules to validate the usefulness of the present approaches. The resulting good correlations between the two data sets demonstrate that the flight characteristics of the propeller-driven airplane can be analyzed effectively through the integrated framework with the propeller and airframe aerodynamic models proposed in this study.