Comparison of traditional and machine learning–based approaches for CIP portfolio selection
| Aspect | Traditional approaches (Simple methods) | Traditional approaches (Complex methods: AHP, mathematical models, fuzzy logic, etc.) | Machine learning–based approach (this study) |
|---|---|---|---|
| Basis of evaluation | Managerial judgment, prioritization matrices, Pareto analysis, cost–benefit | Structured models with multi-criteria decision-making, optimization, fuzzy logic | Historical project data and critical success factors (CSFs) |
| Level of subjectivity | High – depends heavily on experience and personal bias of managers | Moderate – more structured but still requires subjective weightings and expert inputs | Low – relies on data-driven predictions learned from past projects |
| Flexibility and adaptability | Limited – often rigid and not dynamic in fast-changing contexts | Low to moderate – structured but difficult to adapt to changing conditions | High – model adapts as new project data are added (self-updating) |
| Complexity for users | Low – easy to apply but oversimplified | High – requires expertise, time and computational resources | Moderate – requires initial data set preparation but user-friendly once implemented |
| Scalability | Narrow – mainly within one organization or portfolio context | Constrained – scalability limited by data requirements and computational complexity | High – adaptable to different industries, contexts and CSF sets |
| Data requirements | Minimal – often qualitative or subjective inputs | High – requires structured and often large quantitative data sets | Medium – requires historical project data; data augmentation helps overcome small sample issues |
| Transparency | Limited – decisions may appear arbitrary to stakeholders | Moderate – structured, but sometimes opaque in weighting criteria | Moderate to high – model outputs are clear; interpretability tools (e.g. SHAP, LIME) can increase transparency |
| Contribution to organizational learning | Low – decisions not systematically connected to past outcomes | Low – relies on predefined models, not self-learning | High – completed projects continuously update and improve the model |
| Accuracy of project success prediction | Low – prone to inconsistency | Moderate – depends on quality of weights and model calibration | High – especially with algorithms such as multilayer perceptron (MLP), except for underrepresented classes |
| Aspect | Traditional approaches (Simple methods) | Traditional approaches (Complex methods: AHP, mathematical models, fuzzy logic, etc.) | Machine learning–based approach (this study) |
|---|---|---|---|
| Basis of evaluation | Managerial judgment, prioritization matrices, Pareto analysis, cost–benefit | Structured models with multi-criteria decision-making, optimization, fuzzy logic | Historical project data and critical success factors (CSFs) |
| Level of subjectivity | High – depends heavily on experience and personal bias of managers | Moderate – more structured but still requires subjective weightings and expert inputs | Low – relies on data-driven predictions learned from past projects |
| Flexibility and adaptability | Limited – often rigid and not dynamic in fast-changing contexts | Low to moderate – structured but difficult to adapt to changing conditions | High – model adapts as new project data are added (self-updating) |
| Complexity for users | Low – easy to apply but oversimplified | High – requires expertise, time and computational resources | Moderate – requires initial data set preparation but user-friendly once implemented |
| Scalability | Narrow – mainly within one organization or portfolio context | Constrained – scalability limited by data requirements and computational complexity | High – adaptable to different industries, contexts and |
| Data requirements | Minimal – often qualitative or subjective inputs | High – requires structured and often large quantitative data sets | Medium – requires historical project data; data augmentation helps overcome small sample issues |
| Transparency | Limited – decisions may appear arbitrary to stakeholders | Moderate – structured, but sometimes opaque in weighting criteria | Moderate to high – model outputs are clear; interpretability tools (e.g. SHAP, |
| Contribution to organizational learning | Low – decisions not systematically connected to past outcomes | Low – relies on predefined models, not self-learning | High – completed projects continuously update and improve the model |
| Accuracy of project success prediction | Low – prone to inconsistency | Moderate – depends on quality of weights and model calibration | High – especially with algorithms such as multilayer perceptron ( |
Sharing content requires targeting cookies to be enabled. Please update your cookie preferences to use this feature.