Table 4.

Disadvantages in using the Machine Learning/Big Data-AI approaches in research on multimorbidity [14,78,79,80,81,82].

Statements
Typically, require massive data samples for training.
High data quality without missing or biased values.
Enough time for model’s generation in combination of training and testing.
Insufficient prediction performance for clinical practice
Results interpretation, transparency and explainability.
More accurate quantitative measures to evaluate the utility and privacy preservation.
Insufficient validation for clinical practice
High error-susceptibility.