MATLAB: How to plot partial correlation

correlationplot

Hello,

for my analysis I ran a couple of Pearson partial correlations, controlling for variables such as gender and age. Now I want to plot the correlation in a scatterplot. How would I do that? If I use plot(X,Y,'o') then I get only the normal and (not partialled out) correlation, right?

Thank you very much in advance

Best Answer

The partial correlation of X and Y controlling for some other variables [Z] is the correlation of rX with rY, where rX and rY are the residuals from two separate regression equations predicting X from [Z] and predicting Y from [Z] Wikipedia. If you wanted to represent that partial correlation in a plot, you would have to use fitlm to predict X and Y separately, and then get the residuals from those two fits (see Save residuals) as rX and rY. Then you could use

plot(rX,rY,'o');

I'm not sure this plot will really be what you want, though, since the numbers will not be on the same scale as your original X and Y scores (i.e., they will have means of 0). Of course, a plot is just a display device, so maybe this type of display will be fine for you.

A side comment: It probably isn't a good idea to use partial correlation to control for gender as if it were a normally distributed numerical variable, which it isn't. To display the X/Y relationship controlling for gender, it would be better to make separate scatterplots for males versus females or to make a single scatterplot but use separate symbols for the two genders.

Related Solutions

MATLAB: Feature reduction via regression analysis

If you prefer linear regression, use function stepwisefit or its new incarnation LinearModel.stepwise. For example, for backward elimination with an intercept term you can do

load carsmall
X = [Acceleration Cylinders Displacement Horsepower];
y = MPG;
stepwisefit([ones(100,1) X],y,'inmodel',true(1,5))

In general, there is no "best" approach to feature selection. What you can do depends on what assumptions you are willing to make (such as linear model), how many features you have and how much effort you want to invest.

MATLAB: I have prepared 3 D plot for 300 random number with specified std deviation and mean . the question is that how can i draw those points over earth ellipsoid

a=638; % horizontal radius
b=566.3142; % vertical radius
x0=0; % x0,y0 ellipse centre coordinates
y0=0;
t=-pi:0.01:pi;
x=x0+a*cos(t);
y=y0+b*sin(t);
N = 1000 ;
rx = (max(x)-min(x)).*rand(N,1) + min(x); 
ry = (max(y)-min(y)).*rand(N,1) + min(y); 
idx = inpolygon(rx,ry,x,y) ;
figure
hold on
plot(x,y)
plot(rx(idx==1),ry(idx==1),'.r')
plot(rx(idx==0),ry(idx==0),'.b')

Best Answer

Related Solutions

MATLAB: Feature reduction via regression analysis

MATLAB: I have prepared 3 D plot for 300 random number with specified std deviation and mean . the question is that how can i draw those points over earth ellipsoid

Related Question