Hi,
I understand that you want to plot the results of an improved fuzzy c-means (FCM) algorithm and are unsure whether to plot data values or membership values against cluster centers.
I assume you have implemented or have access to the improved FCM algorithm and have the necessary data and membership values available. Here are the steps you can follow:
- Plotting Data Values vs. Membership Values: Typically, when visualizing FCM results, you can plot the data points and use color or markers to indicate membership values relative to each cluster center.
- Using Membership Values: The membership matrix U indicates the degree of belonging of each data point to each cluster. You can use these values to color-code the data points in your plot.
- Centers as a Matrix: In FCM, cluster centers are typically stored as a matrix where each row represents a cluster center, and each column represents a feature dimension. This is a standard representation.
- Plotting Centers: Plot the centers on the same graph to visualize their positions relative to the data points. Use distinct markers to differentiate them from the data points.
- Validation: Ensure that your plot correctly represents the clustering by checking that data points are closer to their respective cluster centers.
Hope this helps!
