Browsing by Subject "Multiple linear regression"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Thumbnail Image
    ItemOpen Access
    Finite computation of the ℓ 1estimator from Huber's M-estimator in linear regression
    (2004) Pınar, M. Ç.
    We review and extend previous work on the approximation of the linear ℓ 1 estimator by the Huber M-estimator based on the algorithms proposed by Clark and Osborne, and Madsen and Nielsen. Although the Madsen-Nielsen algorithm is a promising one, it is guaranteed to terminate finitely under certain assumptions. We describe a variant of the Madsen-Nielsen algorithm to compute the ℓ 1 estimator from the Huber M-estimator in a finite number of steps without any restrictive steps nor assumptions. Summary computational results are given.
  • Thumbnail Image
    ItemOpen Access
    Impact of natural gas price on electricity price forecasting in Turkish day-ahead market
    (Institute of Electrical and Electronics Engineers Inc., 2019) Poyrazoğlu, Oğuzhan Göktuğ; Poyrazoğlu, G.
    The electricity is a regular commodity that is being sold and bought in a highly transparent and efficient market in Turkey. The market is operated by EXIST and an hourly energy price is formed for every hour in the day-ahead market. In Sept. 2018, EXIST also found a central natural gas market in Turkey which enables a ground for all shareholders in the natural gas industry. This study examines the impact of natural gas prices formed in the market on the electricity price. Different predictors are tested to lower the mean absolute percentage error. Addition of past natural gas price into the forecasting model reduces the error from 15.85% to 14.31 % when the average of the last two weeks' natural gas price is used. This may indicate that the current natural gas price affects the electricity market two weeks later. And the final model also reached the 6.67% error for a Saturday, which is a significantly accurate forecast for a volatile electricity market.