ANOVA - analysis of variance functions in QlikView?

[#1]

Are there any ANOVA functions in QLikview, or may I borrow someone's code who has already defined this?

I believe I can get the Degress of Freedom (DF), the summ of squares and the Mean Square model and Mean Square Error, and the F-Ratio in QlikView, with some pre-data setup and definining of calcualtions. But is there an "anova" function that does all of it for ya either in the client-side or scripting functionality? I am also looking for the anova p-value.

Any suggestions?

Stephanie Lind

[#2]

Did you ever get a way of doing the ANOVA in QV?

[ogautier62]

Hello,

Id : identifying field

f : factor

Y : variable

n : sample size

p : count of factor

create a QV sheet with

in dimension :

$(f)

$(id)

in expression :

SST : total sum square

pow($(y) - avg(total $(y)),2)

s2 : variance

sum(   (pow($(y) - avg(total <$(f)> $(y)),2)))
/ (n - count(total distinct $(f)))

residual :

$(y) - avg(total <$(f)> $(y))

SSIntra : sum square intra factor (total colum with row sum)

pow($(y) - avg(total <$(f)> $(y)),2)

SSInter :sum squre inter factor

pow( avg(total <$(f)>$(y)) - avg(total $(y)),2)

r2 (model correlation)= SSinter/s2 = 1 - SSintra/SST

df = p-1, n-p

variances :

Vinter = SSInter/(p-1)

Vintra = SIntra/(n-p)

Fisher stat F = Vinter/s2

data = mtcars R package

y = hp

f = cyl

n = 32

p = 3 (2,4,6)

regards,

[Majed]

This is the only answer I found on Anova test, and it doesn't work on Qlik sense. There is something wrong with the brackets.

It's a pity that something tat can easily be done in Excel, cannot be done straight forward in qlik sense.