%ttesting

%test for significance. 1. This whole bloody process revolves around

%negatives so its easy to get turned around in the naming or meaning. 2.

%I'm including all 3 t-tests because you will get confused, and looking at

%the results of all 3 options saves a lot of time remembering what method

%does what you want. 3. Depending one how strong your results are, you might want a different one anyhow 

%We want to do a one tail test. We only care if the ws is better or equal

%to rt (for now) vs ws being worse than rt. Therefore null hypothesis will

%be that wsAcc >= rtAcc (> comes from the one tail and exclusion).

%Alternative hypothesis will be wsAcc < rtAcc 

%'tail', 'left' Set the alternative hypothesis to be that the population mean of x is less than the population mean of y.

%make x = wsAcc and y = rtAcc

%Accepting the null hypothesis is success (Answer = 0. p >= 0.05)

%Rejecting the null hypothesis is failure (Meaning that wsAcc < rtAcc)

for i = 1:numCategories

    pwsMultiEnsNotGreater(i) =  isNotGreaterTTest(wsMultiEnsPred(:,i), rtEnsPred(:,i), Yaverage(:,i), Yaverage(:,i)); %low p means greater

    pwsMultiEnsNotLess(i) = isNotLessTTest(wsMultiEnsPred(:,i), rtEnsPred(:,i), Yaverage(:,i), Yaverage(:,i)); %low p means less than, high P means not less than (possibly equal)

    pwsMultiEnsNotEqual(i) = isNotSameTTest(wsMultiEnsPred(:,i), rtEnsPred(:,i), Yaverage(:,i), Yaverage(:,i)); %low p means unequal, high p means they're the same

end

wsMultiEnsImprovement = wsMultiEnsSuccess - rtEnsSuccess;

[wsMultiEnsImprovement;pwsMultiEns;pwsMultiEns<0.05];

%T test notes:

%Three types of t tests

%One Sample - If we already have a known (Iron Clad) value from other experiments and

%want to know if this experiment produced a differing result. - Nope

%Paired Two Sample - if you are testing the same guys but after changing

%something. Is this us? Same photos, different processing.

%Independant Two Sample - Probably us?

%a = Yes or no. uselsess

%b = probability they are the same. Lower is better. Less that 0.05 is good

%c = Confidence Interval. 

%d = tstat, the raw T value. df Degrees of Freedom. sd - standar deviation.