Рет қаралды 28
Researchers combine the use of Partial Least Squares Structural Equation Modelling (PLS-SEM) and Artificial Neural Networks (ANNs) to leverage the strengths of both methodologies, providing a comprehensive approach to understanding complex relationships and improving predictive accuracy. Here are the key reasons for this combination:
1. Complementary Strengths:
PLS-SEM: Provides a clear understanding of the structural relationships between latent variables and offers insights into the theoretical model. It is excellent for hypothesis testing, model validation, and understanding the direct, indirect, and total effects within a model.
ANNs: Excel in capturing non-linear relationships and complex patterns in data. They are powerful predictive tools that can handle high-dimensional data and discover hidden relationships that traditional methods might miss.
2. Enhanced Predictive Accuracy:
Combining PLS-SEM and ANNs allows researchers to first use PLS-SEM to identify significant predictors and understand the theoretical model structure. These predictors can then be used as inputs for an ANN to improve predictive accuracy. ANNs can refine and enhance the predictions made by the PLS-SEM model.
3. Handling Complex and Non-linear Relationships:
PLS-SEM is limited in handling non-linear relationships and interactions between variables. ANNs can model complex, non-linear relationships without the need for specifying the functional form of the relationships, providing a more flexible and powerful approach to capturing data complexities.
4. Improved Model Validation:
Using PLS-SEM, researchers can validate the measurement model and ensure the constructs are reliable and valid. Afterward, ANNs can be applied to further validate and test the robustness of the structural model, offering a deeper level of model validation.