I'm trying to build a regression equation to show the relationship between crop yield (Y) and climatic parameters such as; rainfall (RF), minimum temperature (Tmin) and maximum temperature (Tmax) by using data analysis tool in Excel 2013. There it gives summary statistics including, R-squared, standard error, significance F, interception coefficients. So which term in the summary table can be used as the error term (E) in a regression equation as follows;
Y=a+ b*RF + c*Tmax + d*Tmin + E
If that is not calculated within the process of excel, how can I calculate and include error term into the equation formed by excel?? (As per my understanding it doesn't provide any error term)
Or is there any other software which can generate regression equation with error term?
This link may be helpful
Excel Regression Analysis Output Explained
Best luck
https://youtu.be/EZEygGFOTNc
http://www.statisticshowto.com/excel-regression-analysis-output-explained/
https://youtu.be/OlxiOJ26r_k
Enable the function analysis excel data. From then use the Regression tool , it will make a multiple regression of independent variables , thereby generating the statistical t test of your sample . I have helped
You have to recall the fundamentals of multiple regression. In particular,
1. The error term is always generated, just by definition, but some programmes, such as Excel, generate it behind the scenes. Excel is certainly not a programme with which you want to do heavy econometrics -- there are more appropriate statistical packages for that.
2. The error term, by definition, is the difference between the actual value of y and its predicted value. The predicted value, again by definition, is y = beta1 * x1 + beta2 * x2 + ... + betan * xn for that concrete observation with concrete values of y and xs. So if you want to calculate it yourself, you take the actual y values from the relevant range on your worksheet, calculate the predicted y for each observation in a separate range and subtract the predicted y from the actual y. Voila, you got your error term for all observations.
My version of Excel does not do multiple regression. Excel is not usually considered a suitable statistical package.
There are many accepted statistical packages. R, SAS, SPSS, Systat, JMP, Minitab, Statistica, Sigmaplot, and many others. R is widely used, very powerful, but difficult to learn. Documentation is not easy to understand. R is free. SAS is widely used, very powerful, and has good documentation. SAS is expensive. If your university has a site license for SAS, it isn't too bad. It might be best to pick a package where the people around you can give the most help. A friend/teacher/colleague, or just a working group that you can join. I suggest also consulting with someone in the statistics department. If this is a simple linear problem, then you probably don't have quite enough data.
I can see that yield = rainfall. Yet does it not matter if all the rain falls in 4 days time at the beginning of summer with then no rain for months, versus the rain fall is equally distributed across all weeks? Or one could get rain in the growing phase, and then a good rain when the crop is mature and drying. This is not a simple linear regression problem, because you are also interested in the distribution of rain during a season. For this task, Excel will not work. R and SAS will work. The others may work too, but I am less familiar with them.
- Badges
- Science topic