Student's t-test (For Paired samples) calculation

Question

[jatsyebivialet]

Hi everyone,

I'm doing my project on developing a tool for statistical techniques used in micro-array data analysis. For that i selected few techniques such as Student's t-test, P value, Z-Score, and so on. Now i couldn't find a proper algorithm or the steps for calculating t-test for paired samples. So, Please help me in that.

Thank you in advance.

[jatsyebivialet]

Sorry I didn't get you. Are you asking me about the programming language which I'm using.

[Nicolas Rosewick]

google: http://wiki.stat.ucla.edu/socr/index.php/Two_Paired_Sample_T-Test

[jatsyebivialet]

Thank you for that. If possible can you give me the formula for calculating it.

[andrew.j.skelton73]

Here is a formula. By the sounds of things you're more of a CS person (apologies for the blunt assumption), so I'd consider consulting with a statistician to make sure you've got everything right in your code / head.

[jatsyebivialet]

Thank you. I saw that and found it's so useful. Thank you once again.

[statswork0]

This is an excellent article, assumptions and explanation are really so good, All your contributions are very useful for professionals and non-professionals. Thnaks a lot for sharing a awesome article, Keep on posting.

Answer 1

[Nicolas Rosewick]

to compare values x and y in R:

x <- rnorm(10,0,1)
y <- rnorm(10,2,1)
x
 [1] -0.6576916 -0.2038356  0.4096263  1.1366582 -1.0279443 -0.9268312
 [7]  2.0242495 -0.6359770 -0.1778478  0.7623316
y
 [1]  3.4962164  0.6779518  2.1233511  2.8031245 -0.4187195  0.9020173
 [7]  2.0746169  4.3182421  0.4239949  3.4268671
t.test(x,y,paired=T)

    Paired t-test

data:  x and y
t = -3.7919, df = 9, p-value = 0.00427
alternative hypothesis: true difference in means is not equal to 0
95 percent confidence interval:
 -3.0534517 -0.7715332
sample estimates:
mean of the differences 
              -1.912492

[jatsyebivialet]

Thank you for your response. By i'm doing my project in java.

Answer 2

[Michael]

No need to re-implement that, please check for existing libraries and software first:

https://commons.apache.org/proper/commons-math/jacoco/org.apache.commons.math3.stat.inference/index.html

There are also existing complete Java microarray systems:

• Jexpress
• TM4
• Genesis

T-test is not necessarily the best tool for doing DE in microarrays, check for limma, cyberT, SAM.

Finally, microarrays are a dying technology.

[jatsyebivialet]

Thank you