I see MS Excel has several trend-line options; linear, logarithmic, polynomial, exponential, and power functions.
What is the basis / logic of selecting these functions for biological data.
For e.g. I'm interested in understanding change of abundance either transcripts / proteins; my data fitting with polynominal trend-line.
how can I compare different samples in this option?
Interpretation of results is often a guiding principle for model choice. That is why the linear model is the most popular since it is the easiest to interpret. However other models like the polynomial, especially the quadratic, may be chosen if the trend seems to be curvilinear. The log transformation might be chosen to stabilize a non-constant variance, and so on.
The basis should be a theoretical model. If this is lacking, the second best is to chose the simplest model that fits the data reasonably well.
The trendline function in Excel won't allow you to "compare different samples". You will need to specify model that allows to fit different samples (i.e., the model needs coefficients accounting for the possible relevant differences between the groups). Specifying such models in Excel is a apin in the neck. It's possible with LINEST for the usual linear models, and one can use the solver to fit non-linear models. But, really, I woudln_t do that in Excel. A free software withe extremely flexible modelling options is R (www.r-project.org).
You may have a look at these books:
https://www.amazon.de/Applied-Regression-Analysis-Generalized-Linear/dp/1452205663
Chapter Data Analysis Using Regression And Multilevel/Hierarchical Models
Hi Akila,
You can use a model which best fit to your data. For example if your data have exponential trend and we want to use linear trend then we can get misleading results. Look at your data graphs you can get a lot of information to decide which trend line is best for your data.
You can find some basic information from the link below:
http://www.statisticshowto.com/line-of-best-fit/
https://support.office.com/en-ie/article/choosing-the-best-trendline-for-your-data-1bb3c9e7-0280-45b5-9ab0-d0c93161daa8
All the best
- Badges
- Science topic