How can I add noise to linear graph

21 views (last 30 days)
lucky_
lucky_ on 21 Jan 2016
Commented: jgg on 21 Jan 2016
I have 100 sample points and I'm using y=mx+c to create 1d line and add gaussian noise using randn the attached graph is the result that I got when I tried the first code and when I try the second code I get straight line I think there is something wrong with my plot but I'm not sure where ?
code1
sample = 100;
x = linspace(-5,5);
y= 1.6 * x +6;
dataset = randn(sample,2);
figure
plot(dataset,y)
axis([-8 7])
code2
sample = 100;
x = linspace(-5,5);
y= 1.6 * x +6;
dataset = y*randn(sample,2);
figure
plot(dataset)
axis([-8 7]);
this is the result that I want to get

Accepted Answer

jgg
jgg on 21 Jan 2016
Edited: jgg on 21 Jan 2016
You never added the random noise to your dataset:
sample = 100;
x = linspace(-5,5);
y= 1.6 * x +6;
r = randn(2,sample);
figure
plot(x,r(1,:)+y)
hold on
plot(x,r(2,:)+y)
The actual outcome value is y + r(i,:) not just y; you also are plotting it incorrectly before.
  2 Comments
lucky_
lucky_ on 21 Jan 2016
Thanks that worked I'm still new to this so I wasn't sure .. is there function to plot fitted line
jgg
jgg on 21 Jan 2016
Yes. Let's code it up a little more clearly since you want to fit something too:
sample = 100;
x = linspace(-5,5);
y= 1.6 * x +6;
r = randn(2,sample);
y_noise = y + r(1,:);
X = [ones(1,length(x)); x];
beta = regress(y_noise',X');
y_fitted = beta(1) + x*beta(2);
figure
plot(x,y_noise)
hold on
plot(x,y_fitted)
This fits a linear regression curve to the noisy data using the regress function with an intercept. The coefficients are in beta.
If this answer solved your problem, please help others find it by accepting it.

Sign in to comment.

More Answers (0)

Categories

Find more on Descriptive Statistics in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!