Hi scholars,
Myself and a group of young researchers are working on a survey conducted from 2008 to 2017 and we are struggling with Modelling the prevalence of the outcome variables. We are open to anyone who wish to be part of the team who can help out. This person will also serve as co author.
Hi Obasanjo,
One would need more information on the nature of the dataset and what your exact objectives are. That way, it is possible to make suggestions and imagine a possible level of involvement.
.
Before I answer, let me just give brief description about my experience with such a time series data. I was involved with National Cancer Registry data on cancer incidences of 53 cancer sites. The time trend analysis of those cancer sites are attempted and provided in the special report of NCRP. One can go to NCRP site and have access to those reports.
When you have time series data, the first and foremost problem is to decide the model to be used. Very often, linear regression model is fitted and then even forecasting is done for coming years. Please refer to my following publication:
Projections of Cancer Cases in India (2010-2020), Asian Pacific Journal of Cancer Prevention, Vol 11, 2010 - 1045-49.
In case of longer duration data, even you can fit joinpoint regression lines using the joinpoint program. Sometimes, you may need a log transformation or some other transformation before you fit any model.
In case of suspected seasonal variation, you can even go for moving average approach for projection of future values.
In case you need further help, you can contact me on my email: [email protected]
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling