PCA Resolver

Principal component analysis (PCA) is a widely used tool to handle high-dimension data. As typically high-dimension data, grain-size distributions also could be processed by PCA. QGrain has integrated the basic PCA to extract the major variations of GSDs. The interface is very simple:

1. Just click the Load Dataset button to load your GSDs.
2. Select the number of PCs.
3. Click the Perform button to execute the PCA algorithm.
4. Check the explained variances of PCs (the percentages in the legend) to determine whether to add/reduce a component. If the sum of them is less than 90%, it means the number is not enough. If some minor components have very small variances, consider reducing the number.
5. If you are sure the number is appropriate, click the Save button to save the PCA result to an Excel file.