CODEWITHSUNDEEP
matlabrandomdistribution
emre k
emre k

Reputation: 1

how can i take small part(like 100 as a raw data) from normal distribution and change skewness and kurtosis?

x=randn(1,100000);
subplot(3,1,1) , plot(x)
subplot(3,1,2) , histogram(x);
moments={mean(x),std(x),skewness(x),kurtosis(x)};
display(moments)
y=3*x+1;
subplot(3,1,3) , histogram(y);
C=corrcoef(x,y)

I have done this until now.

I produced random numbers with standard normal distribution. Then, I want to plot the small parts and its histogram.

Also, I want to learn how to change skewness and kurtosis.

Upvotes: 0

Views: 80

Answers (2)

Murad Omar
Murad Omar

Reputation: 90

You can do this with a sinh-arcsinh transformation:

see: http://biomet.oxfordjournals.org/content/96/4/761.abstract

Essentially if f(x) is your normal distribution, then replacing x with x1:

x1 = sinh(delta*asinh(x)- epsilon)
y1 = f(x1)

you get another distribution with a different skewness and kurtosis.

Upvotes: 0

Mahsa.Ghasemi
Mahsa.Ghasemi

Reputation: 98

If you want to draw samples from a distribution with some predefined mean, variance, skewness and kurtosis, you can use pearsnd(mu,sigma,skew,kurt,m,n) to create an m-by-n matrix of random variables. See the documentation for more info

http://www.mathworks.com/help/stats/pearsrnd.html

Upvotes: 1

Related Questions