Three Essays on Forecasting in Nonlinear Models PDF Download
Are you looking for read ebook online? Search for your book and save it on your Kindle device, PC, phones or tablets. Download Three Essays on Forecasting in Nonlinear Models PDF full book. Access full book title Three Essays on Forecasting in Nonlinear Models by Scott T. Murdoch. Download full books in PDF and EPUB format.
Author: Scott T. Murdoch
Publisher:
ISBN:
Category : Economic forecasting
Languages : en
Pages :
Get Book Here
Book Description
Author: Scott T. Murdoch
Publisher:
ISBN:
Category : Economic forecasting
Languages : en
Pages :
Get Book Here
Book Description
Author: Young gui Kim
Publisher:
ISBN:
Category : Academic achievement
Languages : en
Pages : 212
Get Book Here
Book Description
Author: Zhengfeng Guo
Publisher:
ISBN:
Category : Econometrics
Languages : en
Pages : 86
Get Book Here
Book Description
Author: Jin-Lung Lin
Publisher:
ISBN:
Category : Econometric models
Languages : en
Pages : 86
Get Book Here
Book Description
Author: Andreas Peter Gottschling
Publisher:
ISBN:
Category : Feedforward control systems
Languages : en
Pages : 284
Get Book Here
Book Description
Author: Mario Giacomazzo
Publisher:
ISBN:
Category : Autoregression (Statistics)
Languages : en
Pages : 187
Get Book Here
Book Description
The primary objective in time series analysis is forecasting. Raw data often exhibits nonstationary behavior: trends, seasonal cycles, and heteroskedasticity. After data is transformed to a weakly stationary process, autoregressive moving average (ARMA) models may capture the remaining temporal dynamics to improve forecasting. Estimation of ARMA can be performed through regressing current values on previous realizations and proxy innovations. The classic paradigm fails when dynamics are nonlinear; in this case, parametric, regime-switching specifications model changes in level, ARMA dynamics, and volatility, using a finite number of latent states. If the states can be identified using past endogenous or exogenous information, a threshold autoregressive (TAR) or logistic smooth transition autoregressive (LSTAR) model may simplify complex nonlinear associations to conditional weakly stationary processes. For ARMA, TAR, and STAR, order parameters quantify the extent past information is associated with the future. Unfortunately, even if model orders are known a priori, the possibility of over-fitting can lead to sub-optimal forecasting performance. By intentionally overestimating these orders, a linear representation of the full model is exploited and Bayesian regularization can be used to achieve sparsity. Global-local shrinkage priors for AR, MA, and exogenous coefficients are adopted to pull posterior means toward 0 without over-shrinking relevant effects. This dissertation introduces, evaluates, and compares Bayesian techniques that automatically perform model selection and coefficient estimation of ARMA, TAR, and STAR models. Multiple Monte Carlo experiments illustrate the accuracy of these methods in finding the "true" data generating process. Practical applications demonstrate their efficacy in forecasting.
Author: 武亮
Publisher:
ISBN:
Category : Economic forecasting
Languages : en
Pages : 306
Get Book Here
Book Description
Author: Jin-Lung Lin
Publisher:
ISBN:
Category : Nonlinear theories
Languages : en
Pages : 148
Get Book Here
Book Description
Author: Dimitris Korompilis-Magkas
Publisher:
ISBN:
Category :
Languages : en
Pages : 0
Get Book Here
Book Description
This thesis explores several aspects of Bayesian model selection in time series forecasting of macroeconomic variables. The contribution is provided in three essays. In the first essay (Chapter 2) I forecast quarterly US inflation based on the generalized Phillips curve using econometric methods which incorporate dynamic model averaging. These methods not only allow for coefficients to change over time, but also for the entire forecasting model to change over time. I find that dynamic model averaging leads to substantial forecasting improvements over simple benchmark regressions and more sophisticated approaches such as those using time varying coefficient models. I also provide evidence on which sets of predictors are relevant for forecasting in each period. In the second essay (Chapter 3) I address the issue of improving the forecasting performance of vector autoregressions (VARs) when the set of available predictors is inconveniently large to handle with methods and diagnostics used in traditional small-scale models. First, I summarize available information from a large dataset into a considerably smaller set of variables through factors estimated using standard principal components. However, even in the case of reducing the dimension of the data the true number of factors may still be large. For that reason I introduce in my analysis simple and efficient Bayesian model selction methods. I conduct model estimation and selection of predictors automatically through a stochastic search variable selection (SSVS) algorithm which requires minimal input by the user. I apply these methods to forecast 8 main U.S. macroeconomic variables using 124 potential predictors. I find improved out of sample fit in high dimensional specifications that would otherwise suffer from the proliferation of parameters. Finally, in the third essay (Chapter 4) I develop methods for automatic selection of variables in forecasting Bayesian vector autoregressions (VARs) using the Gibbs sampler. In particular, I extend the algorithms of Chapter 3 and provide computationally efficient algorithms for stochastic variable selection in generic (linear and nonlinear) VARs. The performance of the proposed variable selection method is assessed in a small Monte Carlo experiment, and in forecasting four short macroeconmic series for the UK using time-varying parameters vector autoregressions (TVP-VARs). I find that restricted models consistently improve upon their unrestricted counterparts in forecasting, showing the merits of variable selection in selecting parsimonious models.
Author: Wanbin Wang
Publisher:
ISBN:
Category :
Languages : en
Pages : 0
Get Book Here
Book Description
In this thesis, we examine the relationship between news and the stock market. Further, we explore methods and build new nonlinear models for forecasting stock price movement and portfolio optimization based on past stock prices and on one type of big data, news items, which are obtained through the RavenPack News Analytics Global Equities editions. The thesis consists of three essays. In Essay 1, we investigate the relationship between news items and stock prices using the artificial neural network (ANN) model. First, we use Granger causality to ascertain how news items affect stock prices. The results show that news volume is not the Granger cause of stock price change; rather, news sentiment is. Second, we test the semi-strong form efficient market hypothesis, whereas most existing research testing efficient market hypothesis focuses on the weak-form version. Our ANN strategies consistently outperform the passive buy-and-hold strategy and this finding is apparently at odds with the notion of the efficient market hypothesis. Finally, using news sentiment analytics from RavenPack Dow Jones News Analytics, we show positive profitability with out-of-sample prediction using the proposed ANN strategies for Google Inc. (NASDAQ: GOOG). In Essay 2, we expand the utility of the information from news volume and news sentiments to encompass portfolio diversification. For the Dow Jones Industrial Average (DJIA) components, we assign different weights to build portfolios according to their weekly news volumes or news sentiments. Our results show that news volume contributes to portfolio variance both in-sample and out-of-sample: positive news sentiment contributes to the portfolio return in-sample, while negative contributes to the portfolio return out-of-sample, which is a consequence of investors overreacting to the news sentiment. Further, we propose a novel approach to portfolio diversification using the k-Nearest Neighbors (kNN) algorithm based on the idea that news sentiment correlates with stock returns. Out-of-sample results indicate that such strategy dominates the benchmark DJIA index portfolio. In Essay 3, we propose a new model called the Combined Markov and Hidden Markov Model (CMHMM), in which observation is affected by a Markov model and an HMM (Hidden Markov Model) model. The three fundamental questions of the CMHMM are discussed. Further, the application of the CMHMM, in which the news sentiment is one observation and the stock return is the other, is discussed. The empirical results of the trading strategy based on the CMHMM show the potential applications of the proposed model in finance. This thesis contributes to the literature in a number of ways. First, it extends the literature on financial applications of nonlinear models. We explore the applications of the ANNs and kNN in the financial market. Besides, the proposed new CMHMM model adheres to the nature of the stock market and has better potential prediction ability. Second, the empirical results from this dissertation contribute to the understanding of the relationship between news and the stock market. For instance, our research found that news volume contributes to the portfolio return and that investors overreact to news sentiment--a phenomenon that has been discussed by other scholars from different angles.
