Forecasting growth with time series models

Article Abstract:

Time series growth can be predicted using various models. The forecasting performance of three such models are compared: regression analysis, a first-order autoregressive, moving average (ARIMA) model and a second-order ARIMA model. The results indicate that regression favors observed growth in the middle of the sample, while first-order ARIMA gives all observed growth periods equal weights. However, second-order ARIMA favors the most recent observed growth, while giving the least importance to early growth periods. A test shows that this model makes the best predictions.

---

Time series analysis supported by power transformations

Article Abstract:

Time series models often benefit from power transformations. However, forecast may need to be retransformed after power transformations which may produce in a bias in the forecast. For the optimal use of power transformations in a time series analysis, two procedures for the selection of the power transformation is presented which minimizes variance. The procedures are applicable to all types of time-series models, easy to use and results are comparable with other methods.

Evaluation, Transformations (Mathematics)

---

Similar abstracts:

• Abstracts: Subset selection of autoregressive time series models. Testing for short term memory in a VARMA process. Monthly data and short-term forecasting: an assessment of monthly data in a VAR model
• Abstracts: Forecasting real-time data allowing for data revisions. Linear combination of restrictions and forecasts in time series analysis
• Abstracts: Forecasting output growth rates and median output growth rates: A hierarchical Bayesian approach. Kernel-based multistep-ahead predictions of the US short-term interest rate
• Abstracts: Power of tests in binary response models: comment. Efficient semiparametric estimation of expectations. A Divisible Search Model of Fiat Money: a comment
• Abstracts: On the revelation of private information in stock market economies. The graphs of the Walras correspondence: the production economies case

This website is not affiliated with document authors or copyright owners. This page is provided for informational purposes only. Unintentional errors are possible.