|
on Econometrics |
| By: | Jonathan B. Hill (Department of Economics, Florida International University) |
| Abstract: | In this paper we develop asymptotically chi-squared tests of extremal serial dependence for heavy-tailed time series, including infinite variance and infinite mean processes. The methodology forms a basis for establishing bivariate tests of extremal dependence usable as a means to test contagion and explosive market co-movements. We use a difference in extreme tail scales as our measure of dependence, which is similar in spirit to the co-difference for stable random processes, and develop left-, right-, and two-tailed portmanteau-type test statistics. Our test statistics have a chi-squared limit distribution under the null hypothesis of independence, or under the hypothesis of extremal white-noise for processes near-epochdependent on a mixing process; and obtain a power of one for extremal dependent processes under general conditions. In a controlled experiment we apply a statistic ranking strategy in order to select a tail fractile nuisance index, resulting in exceptional rejection frequencies under extremal white-noise, ARMA, and asymmetric SETAR hypotheses. We apply one- and two-tailed tests to various macroeconomic and financial time series,and demonstrate low levels of significant, asymmetric (left tail), and persistent extremal dependence in the returns and absolute returns of the NASDAQ, and the absolute returns of the Shanghai Stock Exchange. |
| Keywords: | dependence, white-noise, near-epoch-dependence, mixingales, regular variation, infinite variance, portmanteau test |
| JEL: | C12 C16 C52 |
| Date: | 2005–08 |
| URL: | http://d.repec.org/n?u=RePEc:fiu:wpaper:0513&r=ecm |
| By: | Rajanish Dass |
| Abstract: | Finding frequent patterns from databases has been the most time consuming process in data mining tasks, like association rule mining. Frequent pattern mining in real-time is of increasing thrust in many business applications such as e-commerce, recommender systems, and supply-chain management and group decision support systems, to name a few. A plethora of efficient algorithms have been proposed till date, among which, vertical mining algorithms have been found to be very effective, usually outperforming the horizontal ones. However, with dense datasets, the performances of these algorithms significantly degrade. Moreover, these algorithms are not suited to respond to the real-time need. In this paper, we describe BDFS(b)-diff-sets, an algorithm to perform real-time frequent pattern mining using diff-sets and limited computing resources. Empirical evaluations show that our algorithm can make a fair estimation of the probable frequent patterns and reaches some of the longest frequent patterns much faster than the existing algorithms. |
| Date: | 2005–08–17 |
| URL: | http://d.repec.org/n?u=RePEc:iim:iimawp:2005-08-05&r=ecm |
| By: | Tiemen Woutersen; Jerry Hausman |
| Abstract: | This paper presents a new estimator for the mixed proportional hazard model that allows for a nonparametric baseline hazard and time-varying regressors. In particular, this paper allows for discrete measurement of the durations as happens often in practice. The integrated baseline hazard and all parameters are estimated at regular rate,square root of N , where N is the number of individuals. A hazard model is a natural framework for time-varying regressors if a flow or a transition probability depends on a regressor that changes with time since a hazard model avoids the curse of dimensionality that would arise from interacting the regressors at each point in time with one another. |
| Date: | 2005–08 |
| URL: | http://d.repec.org/n?u=RePEc:jhu:papers:525&r=ecm |
| By: | Ahmed Shamiri (University Kebangsaan Malaysia); Abu Hassan (University Kebangsaan Malaysia) |
| Abstract: | This paper examines and estimate the three GARCH(1,1) models (GARCH, EGARCH and GJR-GARCH) using the daily price data. Two Asian stock indices KLCI and STI are studied using daily data over a 14-years period. The competing Models include GARCH, EGARCH and GJR-GARCH used with three different distributions, Gaussian normal, Student-t, Generalized Error Distribution. The estimation results show that the forecasting performance of asymmetric GARCH Models (GJR-GARCH and EGARCH), especially when fat-tailed asymmetric densities are taken into account in the conditional volatility, is better than symmetric GARCH. Moreover, its found that the AR(1)-GJR model provide the best out-of- sample forecast for the Malaysian stock market, while AR(1)-EGARCH provide a better estimation for the Singaporean stock market. |
| Keywords: | ARCH-Models, Asymmetry, Stock market indices and volatility modeling, SAS/ETS software. |
| JEL: | C1 C2 C3 C4 C5 C8 |
| Date: | 2005–09–08 |
| URL: | http://d.repec.org/n?u=RePEc:wpa:wuwpem:0509015&r=ecm |