Question: Statistical testing for non parametric data
0
gravatar for SOneil
22 months ago by
SOneil40
SOneil40 wrote:

Hi all,

Is there a way to conduct a Kruskal Wallis test or Mann Whitney U on Galaxy Main?

Thank you for your help.

statistics galaxy • 503 views
ADD COMMENTlink modified 22 months ago by Devon Ryan1.9k • written 22 months ago by SOneil40
3
gravatar for Devon Ryan
22 months ago by
Devon Ryan1.9k
Germany
Devon Ryan1.9k wrote:

Your best bet is the jupyter interactive environment. With the python kernel you can:

from scipy.stats import mannwhitneyu

and then use the mannwhitneyu() function.

ADD COMMENTlink written 22 months ago by Devon Ryan1.9k

Thank you for you reply, I have been doing that thus far. Although I am not that experienced in python, but is it possible to create a loop that could read certain datasets from the history and carry out testing in that manner ?

ADD REPLYlink written 22 months ago by SOneil40
2

Certainly! If you post the first few lines of such a dataset I'll show you the appropriate code.

ADD REPLYlink written 22 months ago by Devon Ryan1.9k

Help via Interactive Environments and Devon, how awesome is this!

ADD REPLYlink written 22 months ago by Bjoern Gruening5.1k
1

Something like this should work:

datasets = [2,3,4,5,6] # the dataset numbers of your history    
for data in datasets:
    data = get(data)
    for row in open(data):
         columns = row.split()
         print( mannwhitneyu( columns[] ) ) # here you need to access the columns you want to use
ADD REPLYlink written 22 months ago by Bjoern Gruening5.1k

This is perfect, thank you so much !

ADD REPLYlink written 22 months ago by SOneil40
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 16.09
Traffic: 108 users visited in the last hour