Structural Equation Models (SEM) are widely used to model complex causal relations as the ones defining human behaviors. Several techniques exist to estimate SEM parameters. Among them the PLS Path Modeling (PLS- PM) algorithm is the most widely used technique. In particular, PLS-PM allows taking into account formative blocks of manifest variables. A new way to compute outer weights in the case of formative block of manifest variables has been recently proposed. This approach involves using PLS Regression (PLS-R) in order to compute outer weights even in the case of multicollinearity among the manifest variables of the same block. However, PLS Regression supposes linearity in relations between variables. Following the previous work, we decide to use a non-linear approach to PLS-R in order to estimate measurement model parameters in a non-linear PLS-PM approach to SEM.
RUSSOLILLO, G., TRINCHERA, L. et ESPOSITO VINZI, V. (2009). A Non Linear Regularized Component-based Approach to Structural Equation Modeling. Dans: Statistical Methods for the Analysis of Large Data-sets. CLEUP, pp. 195-198.