Skip to main content

Table 2 Results with minute by minute modality, online recognition

From: Recognition of cooking activities through air quality sensor data for supporting food journaling

 R.F.B.N.N.B.L.R.kNNSVM
TN332,208297,465308,529332,056321,735293,351
FP153936,28225,218169112,01240,396
TP369111,40375192665470710,549
FN12,6264914879813,65211,6105768
Accuracy (%)95.9588.2390.2895.6293.2586.81
Specificity (%)99.5489.1392.4499.4996.4087.90
Recall (%)22.6269.8846.0816.3328.8564.65
Precision (%)70.5723.9122.9761.1828.1520.71
F1-score (%)34.2635.6330.6625.7828.5031.37
  1. Classifiers: Random forest (denoted as R.F., max depth = 10, iterations = 10), Bayes networks (B.N., using K2 hill climbing search algorithm), Naive Bayes (N.B.), k nearest neighbor (kNN, using \(k=1\)), Logistic regression (L.R.), Support Vector Machines (SVM, using polynomial kernel and class balancing)
  2. Numbers in italic indicate the largest value obtained in the experiment for each considered metric