15:3115:31, 30 June 2015diffhist−58
Mann–Whitney U test
If we applied t-test on non normal distribution P(reject H0) can be > 0.05 under H0. This is why t-test cannot be applied on non normal distribution. This is not a question of efficiency.
15:1815:18, 30 June 2015diffhist+358
Wilcoxon signed-rank test
→Test procedure: Seems more logic to speak about exact distribution first rather than the approximation. The esperance of the approximative normal distribution is 0 not 0.5/sigma... Added precision two sided and one-sided test.+Prescision on Siegel
13:2613:26, 20 February 2015diffhist−336
Common spatial pattern
The modification on definition of covariance matrices did not correspond to the definition and add errors on the following of the article (->removed). The added reference did not add knowledge on the method itself but only on application. (->removed)
15:3115:31, 30 June 2015diffhist−58
Mann–Whitney U test
If we applied t-test on non normal distribution P(reject H0) can be > 0.05 under H0. This is why t-test cannot be applied on non normal distribution. This is not a question of efficiency.
15:1815:18, 30 June 2015diffhist+358
Wilcoxon signed-rank test
→Test procedure: Seems more logic to speak about exact distribution first rather than the approximation. The esperance of the approximative normal distribution is 0 not 0.5/sigma... Added precision two sided and one-sided test.+Prescision on Siegel
13:2613:26, 20 February 2015diffhist−336
Common spatial pattern
The modification on definition of covariance matrices did not correspond to the definition and add errors on the following of the article (->removed). The added reference did not add knowledge on the method itself but only on application. (->removed)