• ETRI Journal
• /
• 제18권2호
• /
• pp.87-106
• /
• 1996

This paper proposes a new model as a framework for forecasting demand and technological substitution, which can accommodate different patterns of technological change. This model, which we named, "Adaptive Diffusion Model", is formalized from a conceptual framework that incorporates several underlying factors determining the market demand for technological products. The formulation of this model is given in terms of a period analysis to improve its explanatory power for dynamic processes in the real world, and is described as a continuous form which approximates a discrete derivation of the model. In order to illustrate the applicability and generality of this model, time-series data of the diffusion rates for some typical products in electronics and telecommunications market have been empirically tested. The results show that the model has higher explanatory power than any other existing model for all the products tested in our study. It has been found that this model can provide a framework which is sufficiently robust in forecasting demand and innovation diffusion for various technological products.

• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회 2007년도 추계학술대회 및 정기총회
• /
• pp.183-189
• /
• 2007

In this paper, we developed the fuzzy group method data handling-type (GMDH) Model and applied it to demand forecasting of educational expenses. At present, GMDH family of modeling algorithms discovers the structure of empirical models and it gives only the way to get the most accurate identification and demand forecasts in case of noised and short input sampling. In distinction to fuzzy system, the results are explicit mathematical models, obtained in a relative short time. In this paper, an adaptive learning network is proposed as a kind of fuzzy GMDH. The proposed method can be reinterpreted as a multi-stage fuzzy decision rule which is called as the fuzzy GMDH. The fuzzy GMDH-type networks have several advantages compared with conventional multi-layered GMDH models. Therefore, many types of nonlinear systems can be automatically modeled by using the fuzzy GMDH. A computer program is developed and successful applications are shown in the field of demand forecasting problem of educational expenses with the number of factors considered.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2001년도 ICCAS
• /
• pp.104.1-104
• /
• 2001

This paper presents a technique to development load control and management system in order to limits a maximum load demand and saves electric energy consumption. The computer programming proper load forecasting algorithm associated with programmable logic control and digital power meter through inform of multidrop network RS 485 over the twisted pair, over all are contained in this system. The digital power meter can measure a load data such as V, I, pf, P, Q, kWh, kVarh, etc., to be collected in statistics data convey to data base system on microcomputer and then analyzed a moving linear regression of load to forecast load demand Eventually, the result by forecasting are used for compost of load management and shedding for demand monitoring, Cycling on/off load control, Timer control, and Direct control. In this case can effectively reduce the electric energy consumption cost for 10% ...

• 응용통계연구
• /
• 제22권5호
• /
• pp.1007-1017
• /
• 2009

본 연구는 최근에 활발히 연구가 진행 중인 항공수요 예측 분야에서 사용되는 계절형 ARIMA 모형과 다변량 계절형 시계열 모형과의 성능을 비교한 것이다. 본 연구에서는 국제 여객 수요와 국제 화물 수요 예측을 위하여 실제 자료를 이용하여 비교한 결과 다변량 계절형 시계열 모형이 예측의 정확도 면에서 기존의 일변량 모형보다 우수함을 보였다.

• 한국산업정보학회논문지
• /
• 제18권3호
• /
• pp.35-46
• /
• 2013

국내 전력 수요량 예측을 위한 정확한 분석 모델을 개발하기 위하여 고차원 데이터 군집 분석에 적합한 차원 축소 개념의 부분공간 군집 기법과 SMO 분류 기법을 결합한 전력 수요 패턴 예측 방법을 제안하였다. 전력 수요 패턴 예측은 무선부하감시 데이터 뿐 아니라 소지역 단위의 센서스 정보를 통합하여 시간대별 전력 부하 패턴 분석과 인구통계학 및 지리학적 특성 분석이 가능하다. 서울지역 대상의 센서스 정보 및 전력 부하를 이용한 소지역 전력 수요 패턴 예측 결과 총 18개의 특성 군집을 구성하였으며, 전력 수요 패턴 예측 정확도는 약 85%를 보였다.

• 한국시뮬레이션학회논문지
• /
• 제23권3호
• /
• pp.19-25
• /
• 2014

수요예측은 경영 전략을 포함한 모든 경영 활동의 기초가 된다. 특히 부품의 수요예측은 공급망관리 측면에서 매우 중요한 요소 중 하나이다. 부품의 수요는 다양한 산업에서 종종 간헐적 특성을 포함한다. 간헐적 특성이란 수요가 발생하지 않는 경우가 빈번한 현상을 지칭한다. 간헐적 수요 현상에서는 발생된 수요의 분산이 크고 그 발생간격이 확률적이다. 따라서 간헐적 특성을 갖는 수요를 예측하기 위해서 일반적인 시계열 분석기법이나 인과관계를 이용한 모형(회귀모형)을 사용하는 것은 적합하지 않다. 이는 기존의 방법들이 실제 수요행태를 묘사하기 어렵기 때문이다. 이러한 간헐적 수요의 예측을 위해 마코프 부트스트랩이 개발되었다. 이 방법은 1계차 자기상관성을 반영하며 리드타임 동안 수요의 합이 독립임을 가정하였다. 본 연구에서는 리드타임 내 수요 합의 독립가정을 완화한 부트스트랩 방법을 제안한다. 수정된 부트스트랩 방법에 의해 재추출된 데이터는 실측 데이터의 간헐적 특성을 근사적으로 반영한다. 마지막으로 실측 데이터에 수정된 방법을 적용한 예측 결과를 사례로 제시하고자 한다.

• International Journal of Computer Science & Network Security
• /
• 제23권9호
• /
• pp.1-7
• /
• 2023

Demand prediction is an essential component of any business or supply chain. Large retailers need to keep track of tens of millions of items flows each day to ensure smooth operations and strong margins. The demand prediction is in the epicenter of this planning tornado. For business processes in retail companies that deal with a variety of products with short shelf life and foodstuffs, forecast accuracy is of the utmost importance due to the shifting demand pattern, which is impacted by an environment of dynamic and fast response. All sectors strive to produce the ideal quantity of goods at the ideal time, but for retailers, this issue is especially crucial as they also need to effectively manage perishable inventories. In light of this, this research aims to show how Machine Learning approaches can help with demand forecasting in retail and future sales predictions. This will be done in two steps. One by using historic data and another by using open data of weather conditions, fuel, Consumer Price Index (CPI), holidays, any specific events in that area etc. Several machine learning algorithms were applied and compared using the r-squared and mean absolute percentage error (MAPE) assessment metrics. The suggested method improves the effectiveness and quality of feature selection while using a small number of well-chosen features to increase demand prediction accuracy. The model is tested with a one-year weekly dataset after being trained with a two-year weekly dataset. The results show that the suggested expanded feature selection approach provides a very good MAPE range, a very respectable and encouraging value for anticipating retail demand in retail systems.

• International Journal of Computer Science & Network Security
• /
• 제23권8호
• /
• pp.210-216
• /
• 2023

Demand prediction is an essential component of any business or supply chain. Large retailers need to keep track of tens of millions of items flows each day to ensure smooth operations and strong margins. The demand prediction is in the epicenter of this planning tornado. For business processes in retail companies that deal with a variety of products with short shelf life and foodstuffs, forecast accuracy is of the utmost importance due to the shifting demand pattern, which is impacted by an environment of dynamic and fast response. All sectors strive to produce the ideal quantity of goods at the ideal time, but for retailers, this issue is especially crucial as they also need to effectively manage perishable inventories. In light of this, this research aims to show how Machine Learning approaches can help with demand forecasting in retail and future sales predictions. This will be done in two steps. One by using historic data and another by using open data of weather conditions, fuel, Consumer Price Index (CPI), holidays, any specific events in that area etc. Several machine learning algorithms were applied and compared using the r-squared and mean absolute percentage error (MAPE) assessment metrics. The suggested method improves the effectiveness and quality of feature selection while using a small number of well-chosen features to increase demand prediction accuracy. The model is tested with a one-year weekly dataset after being trained with a two-year weekly dataset. The results show that the suggested expanded feature selection approach provides a very good MAPE range, a very respectable and encouraging value for anticipating retail demand in retail systems.

• 전기학회논문지
• /
• 제57권8호
• /
• pp.1342-1348
• /
• 2008

One important objective of the electricity market is to decrease the price by ensuring stability in the market operation. Interconnected to this is another objective; namely, to realize sustainable consumption of electricity by equitably distributing the effects and benefits of participating in the market among all participants of the industry. One method that can help achieve these objectives is the ^{(R)}$demand-response program, - which allows for active adjustment of the loadage from the demand side in response to the price. The demand-response program requires a customer baseline load (CBL), a criterion of calculating the success of decreases in demand. This study was conducted in order to calculate undistorted CBL by analyzing the correlations between such external or seasonal factors as temperature, humidity, and discomfort indices and the amounts of electricity consumed. The method and findings of this study are accordingly explicated.

• 한국통신학회논문지
• /
• 제24권12A호
• /
• pp.2025-2033
• /
• 1999

본 연구에서는 IMT-2000 서비스의 상용화 실시 시기를 전문가 의견을 바탕으로 하여 예측한 결과 2001년 2월경에 첫 서비스를 실시할 것으로 나타나 전문가들은 대체로 빠른 진척을 예상하고 있음을 알 수 있었다. 또한 본 연구의 가장 중요한 부분인 IMT-2000 서비스 가입 수요예측에서는 가격 경쟁력에 대한 두 가지 경우에 따라 다른 모형을 제시하였다. 본 연구에서는 신제품에 대한 수요예측에 정성적 방법인 전문가 의견법과 정량적 방법인 성장곡선 모형을 결합하여 과거자료가 없는 신제품의 수요예측의 오차를 줄이고자 하였다. 각 성장곡선 모형에 필요한 계수를 전문가들의 주관적인 의견을 근거로 하여 추정하였다.