Comparative analysis of PLS-SEM and XAI techniques applied to various scenarios/factors in business research problems
| Factor | PLS advantages | XAI advantages |
|---|---|---|
| Better insights | - Estimates complex correlations between latent variables - Effective in formative measurement scenarios | - Improved explainability and interpretability - Addresses PLS limitations - Improved accuracy - Emphasis on prediction |
| Small sample sizes and non-normal data distribution | - Handles small sample sizes and non-normal data distributions more effectively than traditional covariance-based techniques | |
| Complex relationships among latent variables | - Effectively estimate complex correlations between latent variables and provide insights into their overall impact on customer satisfaction | - Helps businesses understand co-interactions among variables and identify key drivers of customer satisfaction - Uncovers hidden patterns and intricate relationships that are not apparent using traditional analytical techniques |
| Formative measurement scenarios | - Provides accurate estimates in formative measurement scenarios, where indicators cause the latent variables rather than being caused by them | |
| High interpretability requirement | - Provides additional insights into the importance and contribution of specific factors - Offers clear explanations of factors driving customer satisfaction | |
| Validation and robustness concerns | - Validates results from other analytical methods. - Provides a more reliable assessment of factors influencing airline service quality - Increases confidence in analysis results | |
| Combination for more comprehensive insights | Complementary strengths when applying XAI techniques. - Contributes to improved decision-making - Improves analysis of complex relationships and interactions among variables | Complementary strengths when applying PLS techniques - Contributes to improved decision-making - Improves analysis of complex relationships and interactions among variables |
| Factor | PLS advantages | XAI advantages |
|---|---|---|
| - Estimates complex correlations between latent variables | - Improved explainability and interpretability | |
| - Handles small sample sizes and non-normal data distributions more effectively than traditional covariance-based techniques | ||
| - Effectively estimate complex correlations between latent variables and provide insights into their overall impact on customer satisfaction | - Helps businesses understand co-interactions among variables and identify key drivers of customer satisfaction | |
| - Provides accurate estimates in formative measurement scenarios, where indicators cause the latent variables rather than being caused by them | ||
| - Provides additional insights into the importance and contribution of specific factors | ||
| - Validates results from other analytical methods. | ||
| Complementary strengths when applying XAI techniques. | Complementary strengths when applying PLS techniques |
Source: This information was compiled by the authors from a comprehensive review of related literature, supplemented with rating based on the author’s expertise