Testing independence of two autocorrelated binary time series

被引:2
作者
Chou, Cheng [1 ]
Chu, Chia-Shang J. [1 ]
机构
[1] Peking Univ, China Ctr Econ Res, Natl Sch Dev, Beijing 100871, Peoples R China
关键词
MODELS;
D O I
10.1016/j.spl.2009.09.014
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
Event forecast is a possibly autocorrelated binary time series. A new test for its timing ability is based on the correlation between the discrete autoregressions of the event forecast and the event time series. The new test outperforms the existing market timing tests. (C) 2009 Elsevier B.V. All rights reserved.
引用
收藏
页码:69 / 75
页数:7
相关论文
共 11 条
[1]  
Al-Osh M. A., 1987, Journal of Time Series Analysis, V8, P261, DOI [10.1111/j.1467-9892.1987.tb00438.x, DOI 10.1111/JTSA.1987.8.ISSUE-3]
[2]   Some aspects of the time correlation problem in regard to tests of significance [J].
Bartlett, MS .
JOURNAL OF THE ROYAL STATISTICAL SOCIETY, 1935, 98 :536-543
[3]  
Berchtold A, 2002, STAT SCI, V17, P328
[4]  
CUI Y, 2009, BIOMETRIKA IN PRESS
[5]  
CUI Y, 2009, REVISITING AR UNPUB
[6]   ON MARKET TIMING AND INVESTMENT PERFORMANCE .2. STATISTICAL PROCEDURES FOR EVALUATING FORECASTING SKILLS [J].
HENRIKSSON, RD ;
MERTON, RC .
JOURNAL OF BUSINESS, 1981, 54 (04) :513-533
[7]  
JACOBS PA, 1978, J ROY STAT SOC B MET, V40, P94
[8]   CONDITIONAL LEAST-SQUARES ESTIMATION FOR STOCHASTIC-PROCESSES [J].
KLIMKO, LA ;
NELSON, PI .
ANNALS OF STATISTICS, 1978, 6 (03) :629-642
[9]   SOME ARMA MODELS FOR DEPENDENT SEQUENCES OF POISSON COUNTS [J].
MCKENZIE, E .
ADVANCES IN APPLIED PROBABILITY, 1988, 20 (04) :822-835
[10]  
MCKENZIE E, 1985, WATER RESOUR BULL, V21, P645