|   Source
  • The most common use comparison test between two subset of data

  • For Example try to make hypothesis that left-hand and right-hand have no different batting average
  • t-test is used to approve or decline our null hypothesis, while t-test is parametric test in our statisitic significance test.
  • NULL Hypothesis, is hypothesis for our data to be asked, if both of our data is similar? or the sample acquired through data that drawn from normal distribution(same population).

  • By doing t-test we reduce all the comparison test to one number only, the t value
  • if only contain 1 sample, compare average, to average sample 0
  • if two samples, compare average 0 to average 1.

  • t-value denotes how extreme our results, and probabilty it will reject out null hypothesis

  • the v value denotes how many variable took into t calculations

  • the p value produced by t and v, will then the value that will determine our hypothesis accepted based in our null hypothesis was true
  • suppose we want to know the difference batting avg left and right, giveb our null hypothesis both no different
  • then if pvalue 0.5, the variance between t and v at least 5% at extreme.
  • pcritical acts like some kind of threshold value. and it will determine whether our null hypothesis is accepted
  • this threshold value will be quite tedious if we set it manually. rather, it can be done automatically.

  • by specifyinhg equal-var=true, the calculations will be exactly the same as welch's t-test
  • it will return a tuple, t and p value

  • we can do it by comparing the sample with one another
  • for our null hypothesis set to true, p must be less than p critical

  • Basic implementation Welch's t-test with python