AI-ordlista
Den kompletta ordlistan över AI
Stress Function
Measure quantifying the difference between distances in the original high-dimensional space and those in the reduced-dimensional space, used to evaluate the quality of the MDS representation.
Spatial Configuration
Geometric representation of objects in the reduced-dimensional space, optimized to minimize the stress function and preserve distance relationships.
Shepard Diagram
Scatter plot showing the relationship between original distances and reconstructed distances, used to diagnose the quality of the MDS fit.
Kruskal's Algorithm
Iterative optimization procedure used in non-metric MDS to find the spatial configuration that minimizes Kruskal's stress.
Optimal Scaling
Transformation technique that optimizes the representation of qualitative data as quantitative scores, often used as a complement to MDS.
Correspondence Analysis
Method related to MDS specialized in analyzing contingency tables, visualizing relationships between categorical variables.
Generalized MDS
Extension of classical MDS allowing the use of more flexible distance models and the handling of incomplete or weighted data.
Burt Matrix
Symmetric cross-tabulation used in multiple correspondence analysis, containing cross-frequencies between all pairs of categories of categorical variables.
Coefficient d'Aliasing
Mesure de la déformation introduite lors de la projection en dimension réduite, quantifiant l'écart entre les distances originales et reconstituées.