|
Activity Number:
|
214
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2003 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #300511 |
|
Title:
|
Modeling Menstrual Cycle Length Using a Mixed Distribution
|
|
Author(s):
|
Ying Guo*+ and Amita K. Manatunga and Shande Chen and Michele Marcus
|
|
Companies:
|
Emory University and Emory University and University of North Texas and Emory University
|
|
Address:
|
1231 Clairmont Rd. Apt. 8-B, Decatur, GA, 30030-1233,
|
|
Keywords:
|
menstrual cycle length ; mixed distribution ; kernel density estimation ; optimum cutoff ; conditional probability ; estimating equation
|
|
Abstract:
|
The length of the menstrual cycle is an important indicator of women's ovarian function. The statistical analysis of cycle lengths is complicated by the facts that the distribution of cycle length, featuring a long right tail, is not normal and that variability within and between women is large. We propose a normal and shifted Weibull mixed distribution for the cycle lengths for women with multiple menstrual cycles. The cycle from the normal distribution is considered standard and that from the shifted Weibull distribution nonstandard. The parameters in the proposed distribution are estimated using the maximum likelihood method. The fitted mixed distribution agrees well with the distribution estimated using nonparametric approaches of Kernel estimation and Kaplan-Meier estimation. Based on the mixed distribution, we propose two measures to help determine whether a cycle is standard or nonstandard. Finally, an estimating equation derived from the score function of the mixed distribution is applied to simultaneously modeling the effect of woman's age on the mean and variation of both the standard and nonstandard cycle lengths.
|
