Skip to main content

Table 10 Classification Report: Evaluation results for Problem (b) (Predict No Dementia, Minimal or Mild Dementia and Moderate or Severe Dementia) using a decision-tree model, in terms of precision, recall and f1-score

From: Identifying the presence and severity of dementia by applying interpretable machine learning techniques on structured clinical records

 

Decision tree model

Precision

Recall

f1-score

Support

No-Dementia

0.91

0.97

0.94

549

Minimal or Mild Dementia

0.81

0.78

0.79

384

Moderate or Severe Dementia

0.74

0.65

0.69

141

macro avg

0.82

0.80

0.81

1074

weighted avg

0.85

0.86

0.85

1074

  1. The support represents the number of true instances of each class. The macro average and weighted average calculate the metrics for each class label. However, the macro illustrates the unweighted mean, without considering label imbalance, whereas the weighted average utilises the support of labels for producing the weighted mean value