Browsing by Subject "EM Algorithm"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Thumbnail Image
    ItemOpen Access
    GDP nowcasting using high frequency asset price, commodity price and banking data
    (2011) Balkan, Binnur
    Knowing the current state of the economy is important especially when we consider that GDP information comes with a lag of quarter. From this perspective, employing high frequency variables in GDP nowcasting may contribute to our knowledge of economic conditions, since they are timelier compared to GDP. This paper deals with nowcasting US GDP using an expectation maximization algorithm in a Kalman Ölter estimation, which includes asset prices, commodity prices and banking data as explanatory variables together with real variables and price indices. As a result of the estimations, asset prices and other high frequency variables are found useful in nowcasting US GDP contrary to previous studies. Model predictions beat the traditional methods with the medium size model, which includes Öfteen variables, yielding the best nowcast results. Finally, this paper also proposes a new route for achieving better nowcast results by changing system speciÖcations of the state variables.
  • Thumbnail Image
    ItemOpen Access
    Maximum likelihood estimation of parameters of superimposed signals by using tree-structured EM algorithm
    (1997) Pakin, Sait Kubilay
    As an extension to the conventional EM algorithm, tree-structured EM algorithm is proposed for the ML estimation of parameters of superimposed signals. For the special case of superimposed signals in Gaussian noise, the IQML algorithm of Breşler and Macovski [19] is incorporated to the M-step of the EM based algorithms resulting in more efficient and reliable maximization. Based on simulations, it is observed that TSEM converges significantly faster than EM, but it is more sensitive to the initial parameter estimates. Hybrid-EM algorithm, which performs a few EM iterations prior to the TSEM iterations, is proposed to capture the desired features of both the EM and TSEM algorithms. Based on simulations, it is found that Hybrid-EM algorithm has significantly more robust convergence than both the EM and TSEM algorithms.