In many psychological studies variables are measured for some subjects in different conditions. In these cases the available information is stored in a three-way data array. Three-way extensions of Principal Component Analysis have been introduced to summarize such an array through components. Among them a widely used method is the so-called Candecomp/Parafac. Its applicability may be limited by the risk of obtaining degenerate solutions (i.e. solutions with highly collinear, diverging and therefore uninterpretable components). This work focuses on some remedies for avoiding degeneracy based on the use of (hard and/or soft) orthogonality constraints.
Giordani, P., Rocci, R. (2016). Remedies for degeneracy in Candecomp/Parafac. In Quantitative Psychology Research (pp. 213-227). Springer New York LLC [10.1007/978-3-319-38759-8_16].
Remedies for degeneracy in Candecomp/Parafac
ROCCI, ROBERTO
2016-01-01
Abstract
In many psychological studies variables are measured for some subjects in different conditions. In these cases the available information is stored in a three-way data array. Three-way extensions of Principal Component Analysis have been introduced to summarize such an array through components. Among them a widely used method is the so-called Candecomp/Parafac. Its applicability may be limited by the risk of obtaining degenerate solutions (i.e. solutions with highly collinear, diverging and therefore uninterpretable components). This work focuses on some remedies for avoiding degeneracy based on the use of (hard and/or soft) orthogonality constraints.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
