t-test for non-equal distribution variances

[daryadedik]

Current: for t-test we assume that two populations have equal variances. The case of non-equal variances is not implemented.
Possible solution: add Welch's t-test implementation for the case of unequal populations variances.

Note: this question should be preliminary investigated: when it's necessary or helpful, when it's not.