The aim of the present study is to build some
mathematical models and then to forecast some fertility parameters
in urban area of Bangladesh. For this purpose, the secondary time
series data on Crude Birth Rate (CBR), Total Fertility Rate (FFR).
Gross Reproduction Rate (GRR) and Net Reproduction Rate (NRR)
of various issues duly published by Bangladesh Bureau of Statistics
(BBS) have been used in the present study. A few mathematical time
trend models have been fitted to time series data of CBR, TFR, GRR
and NRR It is _found that the CBR follows quadratic H.e. parabolic)
polynomial model while the TPR, GRR and NRR follow simple
linear regression model. Model validation technique .such as Cross-
Validity Prediction Power (C VFP), pi, , is applied to these models
to verify how much these models are valid or not. It was found that
all these models are more than 95%, 79%, 82%, and 72% stable
respectively and their shrinkages are only 0.00739Z 0.032l33.
0. 027916, and 0.0424229, respectively. These rates have been
forecasted during 1999-2005 using these time trend models.
