Here you can learn about standard deviation that is called parameter!

 

A standard deviation that is called parameter estimates a test which is making use of mean and standard deviation as part of the calculation are parametric test a test which is not dependent on the mean and standard deviation is called nonparametric test okay typically in a nonparametric test what happens is you arrange you convert the numbers into ranks and then based on a ranking position you do the test let's hope nonparametric test works so no details you can go and search but I what I am trying to say is that data point is converted for example schools of the student in that case or is an absolute number.

What I can do is based upon the score I can see this is Rank 1 this is rank 2 this is rank 3 ranked for rank 5 in the class and then I used before rank after rank and then do the test compare the results those kind of tests are called as non parametric tests well in that case you are now not dependent on mean and standard deviation okay so we have completed t-test let's quickly take a break all of you I hope you're able to follow lots of statistics is being bombarded I understand all of that you may not completely understand in one go you have to see the beat video and this book I have seen that this book is really good so do refer the book try to solve some problems I would advise that solve at least two to three problems on your own all solve the solve problems on your own and see the solution once you do that that is sufficient to understand when I was noting here I wrote here failed to reject H naught and I said meaning accept alternate we are saying fail to reject null meaning except now sorry for that and null hypothesis in Shapiro is basically its normal distributed and reject H naught meaning except I tried okay so the so this is a correction.

When you see the video the second part you have corrected okay is that okay so let's move on to our next stats test which is ANOVA see in a t-test see in a t-test what you can test is one sample concerning respect to a reference point or you then have two samples or you can have two samples which are paired but the moment you have to go from two samples to three samples or more than three samples you cannot use a t-test your t-test is maxed for two samples and what does t-test prove whether that means of the two samples are equal or not equal that's what Betty testers if you have more than two samples then you have to go to ANOVA which is an analysis of variance okay this is what I've said t-test compares mean between two samples what if there are more than two samples and that's where ANOVA comes into play what we are going to discuss is one-way.

ANOVA which means we are just changing one parameter to test whether there is an impact or not not sure but why I feel that during my school days we had this uh no odd thing taught or probably junior college it was taught and the time we had this one way ANOVA probably you would have also been learning this and there would be one example which typically is given to teaching ANOVA concept and that example is typically given is and this is also used by the way in the design of experiment the extension of this design of experiments say you've got this form you have this farm broken into many elements now I know you are trying to test fertilizer has gots a composition ABC these different compositions so you say this is a 68 this is 20.

So you may have one experimentation happening here with some combination then you may have different composition happening you try that 6040 zero something and you have various combinations and then you try to see where the yield is coming higher okay and based upon where you get a higher yield you try to find out which composition because you are not you cannot try all possible component composition basically you cannot try 59 19 and 22 you cannot try that you have some combinations you can do enter try it out and then based upon the various EAL outcomes you determine what is the optimum combination of the three that is entirely something or design of experiments create various combinations said that by using all the outcomes of different combinations basically if this is a grid of six by four 24 so here now you have got 24 outcomes coming and then you will have 24 equations right because there is a a b c equal to yield right and then using those combinations you may solve the equation and then find what is the optimal of a B and C these are all optimization problems these are all optimization problems which are part of design of experiments.

We are going to take a very simple approach which is a simple one way ANOVA okay which is an analysis of visions here only one parameter is changing you only one parameter is changing so you know one way or no, ah what do you have is simply I won't say fertilizer X fertilizer Y fertilizer said ok and here probably many experiments are happening so that finally I get the mean standard deviation of this here many experiments are happening I get mean standard deviation I hear many experiments happening mean standard deviation once I have got the mean and standard deviation that is variance square of standard deviation mean and variance.

What then I will have to prove whether there is a difference in mean of these groups whether there is a difference or there is no difference ok null hypothesis are all means are equal null hypothesis in unwise all means are equal which means the outcome of fertilizer X outcome of fertilizer y outcome of fertilizer is it alternate hypothesis is mean are at least one of the mean is not same alternate here is not all means are not same the null hypothesis alternate hypothesis is framed as at least one of the mean is not equal are you ready so we are in a one way ANOVA if you change more than one parameter it becomes two way that is manova multi-way ANOVA ok but for us one way is good enough to learn and then depending upon your problem statement as and when you face you may go into the multi wmulti-wayl of those things and design of experiments null hypothesis alternate hypothesis which I disagree is that clear.

Yeah one-way ANOVA is an only bus test statistics what it means is when I apply an ANOVA it will prove either all means are same or all means are not same but it cannot tell me which of the mean is not same it will not give me that outcome because the end of the day when you run this it is going to throw a p-value it's going to throw a p-value and that p-value interpretation is at least one of the mean is not same which one is not saying which one is higher which one is lower are you compute the mean and get that outcome right or there is a test which you can rerun to find between pairs which of the means are not significantly same or different okay to determine which specific group are significantly different from each other you have to apply a post hoc test ok.

So if you want that further that is called tusky I I don't recollect tusky test something let me just see yeah tusky a turkey HSD test that is a post hoc test you have to apply on the same data to then get which of the means which of the pair whether fertilizer X versus Y fatty like the X versus a what is a virus is that where there is a significant difference if you want that information then you have to apply that to key SSD test ok.