Independent variables or not?

Hi,

I have a set of parameters and I create a set of observations varying each parameter with a 50% step (E.g. param1 = 0.0; 0.5; 1). I would like to know if these parameters are dependent on each other or not. They have an influence on what I have to measure, but when I compute the correlation matrix I observe that correlation (param xi, param xj) = 0 and I conclude that they are independent. I am right or wrong? Please help me. Thanks a lot!

Elena

Answer 1

[Walter Roberson]

I suggest you give the command

format long g

and then display the correlation matrix again.

[Hotbirdym]

Hi,

Thank you for your answer. The value is something like 1*pow(10, -20) which is very low. It is correct to say that these variables are independent? Though, the correlation between each parameter and my measure starts from -0.20 until -0.60 and I think that as one increases the other decreases.

If the parameters are independent and the only dependent variable is my measure, then I will use MLR to determine the weight of each parameter in order to compute my measure. If they are dependent, then I will use PCA.

I've tried both of the methods, but I obtain strange results and I don't know what am I doing wrong. Please give me a suggestion. Thanks!