How to calculate scores when using pcacov?

I want to perform PCA analysis. When using pcacov function, it is not returned the scores. Could any of you guys help me with this problem? I know that using pca function scores are returned. But I am not using pca function because I want to perform PCA analysis using the Spearman correlation matrix obtained instead of Pearson. Thanks a lot for your help.

Answer 1

[arushi]

Hi Juan,

To perform PCA using a Spearman correlation matrix and obtain scores, you'll need to follow a few steps, as the pcacov function in MATLAB does not directly return scores. Here's how you can accomplish this:Steps to Perform PCA with Spearman Correlation

Compute the Spearman Correlation Matrix:

• First, calculate the Spearman correlation matrix for your data. You can use the corr function with the 'Spearman' option:

    data = [...]; % Your data matrix
    spearmanCorr = corr(data, 'Type', 'Spearman')

Perform PCA Using pcacov:

• Use the pcacov function to perform PCA on the Spearman correlation matrix. This function will return the eigenvectors (coefficients) and eigenvalues, among other outputs.

Compute the Scores Manually:

• To obtain the scores, you need to project your standardized data onto the principal component space. First, standardize your data (mean = 0, variance = 1), and then multiply by the coefficients:

    [coeff, latent, explained] = pcacov(spearmanCorr);
    standardizedData = zscore(data);
    scores = standardizedData * coeff;

Hope this helps.