IFS Class Geometry Explorer v0.8.9

Class-structured fully and partially specified data in the intuitionistic fuzzy triangle.

Case study

Expected columns: ID, mu, nu, Class; optional: Year. Leave nu blank for a μ-only observation or mu blank for a ν-only observation.

Add point / partial profile

Classes

Classshowpointsμ onlyν onlycolor

Turning a class back on resets all three subtypes to visible.

Layers

labels
avoid label overlap
leader lines for displaced labels
point centroids
point medoids
convex hulls of exact points
convex hulls of all admissible objects

min–max bounding rectangles

central region
central coverage circle

Both layers use the same user-defined percentage of exact points and the same center. The circle radius is the distance to the farthest selected central point.

concave hull / alpha shape

α controls tightness: higher α → tighter / more concave boundary.

class-hull intersections

Uses convex hulls of exact points. Empty intersections are reported explicitly; non-empty intersections are filled in purple. Jaccard and containment coefficients appear in statistics.

partial-segment intersections with class regions

Highlights the portions of μ-only / ν-only admissible segments that lie inside exact-point class hulls.

centroid bisectors

Centroids are computed from exact points only.

temporal trajectories

No temporal repeated IDs in current dataset.

Visual convention

full specification: exact point \((\mu,\nu)\)
partial specification: μ-only / ν-only segment
exact-point centroid (darker class color)
exact-point medoid (darker class color)
exact-point convex hull
clipped min–max rectangle
central region
central coverage circle
alpha shape

Selected profile

Click a point or segment.

Class statistics

Pairwise overlap

Enable class-hull intersections.

Profiles

ID	μ	ν	π	Specification	Class

Medoid: observed exact point minimizing the total Euclidean distance to all other exact points in its class. Central region: the user-selected fraction of exact points closest to the chosen center. Alpha shape: a Delaunay/alpha boundary; it may be disconnected when α is large. Analysis/rendering separation: analytical layers use all data of an active class; subtype checkboxes control visibility only.