File Exchange > Data Analysis >    Sparse Principal Components Analysis

Author:
OriginLab Technical Support
Date Added:
4/4/2023
Last Update:
1/9/2024
Downloads (90 Days):
919
Total Ratings:
6
File Size:
446 KB
Average Rating:
File Name:
Sparse_Pri...is.opx
File Version:
1.20
Minimum Versions:
License:
Type:
App
Summary:

Find sparse principal components for optimally reconstructing data.

Screen Shot and Video:
Description:

Purpose

This app is capable of finding the sparse principal components for optimally reconstructing data.

Features include:

  • Support both Sparse PCA and Mini-Batch Sparse PCA.
  • Group support in Transformed Data Plot and Biplot.
  • Confidence ellipse in Transformed Data Plot and Biplot.
  • 3D plot support for Component PlotTransformed Data Plot and Biplot.
  • Outlier detection in Transformed Data Plot and Biplot.

Installation

Download the file "Sparse Principal Components Analysis.opx", and then drag-and-drop onto the Origin workspace. An icon will appear in the Apps gallery window.

NOTE: This tool requires OriginPro.

Operation

Activate a worksheet with data selected,

  1. Click the Sparse Principal Components Analysis icon in the Apps Gallery window to open the dialog.
  2. In the Input tab, choose data in the worksheet for Input Data, where each column represents a variable.
    You can also choose a column for Observations, which can be used for labels in Transformed Data Plot and Biplot.
    Group can be used to divide observations in Transformed Data Plot and Biplot.
  3. In the Settings tab, Mini Batch option determines whether to perform Mini-Batch Sparse PCA or not. 
    Number of Components to Extract is used to control output of components, transformed data and their plots.
    Please refer to SparsePCA and MiniBatchSparsePCA for more details.
  4. In Quantities to Compute tab, check options to control which results to output in Report Data sheet.
  5. In Plots tab, specify whether to create Error Plot (for Mini Batch is not checked)Component Plot, Transformed Data Plot and Biplot.
    All except Error Plot support 2D and 3D. The last two also support confidence ellipse and labeling of outliers.
  6. Click OK button. A report sheet, a report data sheet and a plot data sheet will be created.
    If Show Confidence Ellipse option is checked in Plots tab, a Matrix book will also be created.

Algorithm

The algorithm for this app is from the 3rd Python library, scikit-learn. For more details, please refer to this page.

Updates:

6/6/2023, v1.01, correct category
6/29/2023, v1.1, support for adjusted variance

Reviews and Comments:
11/30/2023101121看看就

09/06/2023wonice问问

09/06/2023wonice

08/12/2023SerbianNSEcellent

06/19/202313129986546no bug reports

06/09/2023wzc9921