Correlation Matrix Calculation | Vose Software

Correlation Matrix Calculation

See also: ModelRisk functions and windows, Modeling correlation introduction, Correlation in ModelRisk

 

Introduction

The Correlation Matrix Calculation window calculates and visualizes the rank order correlation matrix of a data set.

The correlation matrix contains Spearman's rank order coefficient (also known as rho) for each pair of datasets.

It is symmetric because correlation between A and B is the same as correlation between B and A. It's elements lie in the [-1,1] interval.

As the correlation of a variable with itself is 1, the diagonal elements of the matrix are all equal to 1.

The output of the function is an nxn array where n is the number of variables in the data set.

To see the output functions of this window, click here.

Window Elements

In the location field you can indicate the range of spreadsheet cells that contain the data. You can specify whether these are orientated in columns (selected by default, as this is usually the case) or rows. The number of columns (respectively rows) is the n mentioned above.

The correlation matrix of the data is shown. Its elements are the pairwise correlation of each of the variables within the dataset.

Optionally, in the Labels field you can specify where the labels of the dataset are in the spreadsheet. If no labels are selected, the datasets will be named Var1, Var2, etc..

In the Output location field you can specify where the correlation matrix should be placed in the spreadsheet. It will be inserted there upon pressing the OK button.

The number of data values for each variable (source points) is displayed below the graphs.

Selecting any white square in the correlation matrix will generate a scatter plot of the data for the corresponding two variables. You can choose whether to display the actual data values or the percentiles. In the first case, the horizontal and vertical axes are adjusted to the range of the data. When showing percentiles, the range of the axes is [0,1].

 

Navigation