Related Experiment Videos
Adjustment of provisional mortality series: the dynamic linear model with structured measurement errors
Journal of the American Statistical Association
|September 1, 1991
Summary
This study introduces a bivariate structural model to adjust provisional time series data, accounting for correlated measurement errors. The method improves forecasting accuracy for time series analysis, demonstrated with mortality data.
Area of Science:
- Statistics
- Biostatistics
- Epidemiology
Background:
- Provisional time series data often contain errors that affect accuracy.
- Adjusting these series is crucial for reliable statistical analysis and forecasting.
- Existing methods may not adequately address correlated measurement errors.
Purpose of the Study:
- To develop a bivariate structural model for adjusting provisional time series.
- To incorporate correlated measurement errors within the time series model.
- To provide a minimum mean squared error adjustment procedure for provisional and final series.
Main Methods:
- A bivariate structural model with common trend and seasonal components was employed.
- Maximum likelihood estimators were derived for model parameters.
- A minimum mean squared error (MMSE) adjustment procedure was developed.
Main Results:
- The model effectively adjusts provisional time series by accounting for common and specific error structures.
- The MMSE procedure provides an optimal adjustment under the specified model.
- The technique was successfully illustrated using provisional data for forecasting ischemic heart disease mortality.
Conclusions:
- The proposed bivariate structural model offers a robust method for adjusting provisional time series with correlated errors.
- This approach enhances the reliability of statistical inferences and forecasts derived from provisional data.
- The technique has practical applications in public health surveillance and mortality forecasting.