Impact of C07 on weights
| Item ID | Dimension–Pillar and item focus | Effect on cluster percentiles (α) | Effect on entropy (β) | Effect on SHAP relevance (γ) | Implication for |
|---|---|---|---|---|---|
| ID3 | D-PM: use of IoT devices and real-time sensors for equipment health monitoring | Widens the gap between upper and lower quantiles due to C07’s low adoption | High dispersion preserved | High SHAP, as it strongly contributes to the Digital deficit | Potential weight increase (emerging differentiator) |
| ID5 | G-MP: energy monitoring and improvement actions in maintenance operations | Reinforces a low lower-quartile, confirming systematic underperformance | Moderate entropy (stable weakness across firms) | Moderate SHAP as a persistent negative driver | Slight weight increase (still discriminative) |
| ID12 | D-OTM: cross-functional use of digital tools in operations and teamwork | Convergence of scores around the upper quantiles | Entropy decreases (practice becoming standard) | Low SHAP, as it no longer explains maturity differences | Likely weight decrease (no longer informative) |
| ID18 | G-FI: integration of eco-efficiency criteria in inspection and quality routines | Increases heterogeneity in the central and upper quantiles | Entropy increases (diverging adoption patterns) | High SHAP for explaining the Green deficit | Potential weight increase (key indicator of G maturity) |
| ID22 | H-T&E: frequency and coverage of operator training and education in maintenance-related tasks | Upper quantiles further consolidated at high values | Low entropy (uniform good practice across the cluster) | Low SHAP, as it rarely drives maturity differences | Likely weight decrease (mature, non-discriminative item) |
| Item ID | Dimension–Pillar and item focus | Effect on cluster percentiles (α) | Effect on entropy (β) | Effect on SHAP relevance (γ) | Implication for |
|---|---|---|---|---|---|
| ID3 | D-PM: use of IoT devices and real-time sensors for equipment health monitoring | Widens the gap between upper and lower quantiles due to C07’s low adoption | High dispersion preserved | High SHAP, as it strongly contributes to the Digital deficit | Potential weight increase (emerging differentiator) |
| ID5 | G-MP: energy monitoring and improvement actions in maintenance operations | Reinforces a low lower-quartile, confirming systematic underperformance | Moderate entropy (stable weakness across firms) | Moderate SHAP as a persistent negative driver | Slight weight increase (still discriminative) |
| ID12 | D-OTM: cross-functional use of digital tools in operations and teamwork | Convergence of scores around the upper quantiles | Entropy decreases (practice becoming standard) | Low SHAP, as it no longer explains maturity differences | Likely weight decrease (no longer informative) |
| ID18 | G-FI: integration of eco-efficiency criteria in inspection and quality routines | Increases heterogeneity in the central and upper quantiles | Entropy increases (diverging adoption patterns) | High SHAP for explaining the Green deficit | Potential weight increase (key indicator of G maturity) |
| ID22 | H-T&E: frequency and coverage of operator training and education in maintenance-related tasks | Upper quantiles further consolidated at high values | Low entropy (uniform good practice across the cluster) | Low SHAP, as it rarely drives maturity differences | Likely weight decrease (mature, non-discriminative item) |