Calculating Average of Principle Component Analysis (PCA) Data

Hello, I have just performed PCA on my data of sensor values that are infive different groups in an excel file. So now I would like to calculate the average of each group of PCA points. I'm having trouble understanding how the PCA data isstored and then the process of going throguh it to setup five different averages. Thanks so much to whoever can help!

-Zack

Answer 1

[Aditya]

Hi Zack,

To calculate the average of PCA-transformed data for each group, you'll need to follow these steps:

1. Perform PCA.
2. Organize Your Data by Groups
3. Calculate Group Averages

% Example of performing PCA
% dataMatrix is your original data matrix
[coeff, score, ~, ~, explained] = pca(dataMatrix);
% Read data and group labels from Excel
[numData, txtData, rawData] = xlsread('yourfile.xlsx');

% Assuming the last column contains group labels
groupLabels = rawData(:, end);
dataValues = numData; % Assuming the numeric data is in the rest of the columns
% Unique group identifiers
uniqueGroups = unique(groupLabels);

% Initialize a matrix to store the average PCA scores for each group
numComponents = size(score, 2); % Number of principal components
groupAverages = zeros(length(uniqueGroups), numComponents);

% Calculate the average for each group
for i = 1:length(uniqueGroups)
    group = uniqueGroups{i};
    groupIndices = strcmp(groupLabels, group);
    groupScores = score(groupIndices, :);
    groupAverages(i, :) = mean(groupScores, 1);
end

% Display the group averages
disp('Group Averages for PCA Scores:');
for i = 1:length(uniqueGroups)
    fprintf('Group %s: %s\n', uniqueGroups{i}, mat2str(groupAverages(i, :)));
end