Get means for nested categorical variables

1 次查看（过去 30 天）
Anthony
Anthony 2018-3-13
回答： Razvan Carbunescu 2018-3-16
Assuming you have the following table:
site = {'1','1','1','1','2','2','2','2','3','3','3','3'}';
site = categorical(site);
days = {'0','0','45','45','0','0','36','36','0','0','67','67'}';
days = categorical(days);
var = [23,45,67,21,42,87,84,65,12,77,42,90]';
T = table(site,days,var);
how do I make a new table T1 with each unique value of site and days, and the means of var over the days category? e.g. T1 = table(site,days,meanvar)

请先登录，再回答此问题。

采纳的回答

Akira Agata
Akira Agata 2018-3-14
Please try the following:
[group,site,days] = findgroups(T.site,T.days);
meanvar = splitapply(@mean,T.var,group);
T1 = table(site,days,meanvar);

更多回答（2 个）

Peter Perkins
Peter Perkins 2018-3-15
Another possibility:
>> T2 = varfun(@mean,T,'GroupingVariables',{'site' 'days'})
T2 =
6×4 table
site days GroupCount mean_var
____ ____ __________ ________
1 0 2 34
1 45 2 44
2 0 2 64.5
2 36 2 74.5
3 0 2 44.5
3 67 2 66
By the way, "1", "2", etc. are perhaps perfectly good category names for the site variable (as long as you remember that the values are text, and not the numeric values 1, 2, etc), but it seems like days is a numeric count. If that's true, there's no particular reason why you can't just leave it as numeric. The above code, and indeed the version usig findgroups/splitapply, continues to work.
Razvan Carbunescu
Razvan Carbunescu 2018-3-16
Can also use groupsummary to compute directly one go from T:
>> GT = groupsummary(T,{'site','days'},'mean','var')
GT =
6×4 table
site days GroupCount mean_var
____ ____ __________ ________
1 0 2 34
1 45 2 44
2 0 2 64.5
2 36 2 74.5
3 0 2 44.5
3 67 2 66

类别

在 Help Center 和 File Exchange 中查找有关 NaNs 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by