Cough detection (precision and recall)
24 hours
Measure C-mo System's performance and ability to automatically detect cough, using precision and recall (percentage - between 0% and 100%), higher scores mean a better outcome.
Cough detection (F1-score)
24 hours
Measure C-mo System's performance and ability to automatically detect cough, using the F1-score (value between 0 and 1), higher scores mean a better outcome.
Cough characterisation (precision, recall and global accuracy)
24 hours
Measure C-mo System's performance and ability to automatically characterise cough, using precision, recall, and global accuracy (percentage - between 0% and 100%), higher scores mean a better outcome.
Cough characterisation (F1-score)
24 hours
Measure C-mo System's performance and ability to automatically characterise cough, using the F1-score (value between 0 and 1), higher scores mean a better outcome.
Cough characterisation (Matthews correlation coefficient)
24 hours
Measure C-mo System's performance and ability to automatically characterise cough using Matthews correlation coefficient (MCC). The MCC value ranges from -1 to 1, and in this case it indicates the level of agreement between C-mo System's output and the result obtained from expert analysis (considered to be the gold standard in this study). A MCC value of -1 indicates total disagreement, 0 indicates that C-mo System's classification is no better than random guessing, and 1 represents a perfect classification (total agreement between C-mo System's output and the gold standard). Hence, higher scores mean a better outcome.
Cough characterisation (Cohen's Kappa)
24 hours
Measure C-mo System's performance and ability to automatically characterise cough using Cohen's Kappa coefficient (κ). The κ value indicates the level of agreement between C-mo System's output and the result obtained from expert analysis (considered to be the gold standard in this study). It ranges from -1 (worst possible performance) to 1 (best possible performance). Hence, higher scores mean a better outcome.
Wheezing detection (precision, recall, true negative rate, accuracy, and negative predictive value)
24 hours
Measure C-mo System's performance and ability to automatically detect wheezing in cough events, using precision, recall, true negative rate, accuracy, and negative predictive value (percentage - between 0% and 100%), higher scores mean a better outcome.
Wheezing detection (F1-score)
24 hours
Measure C-mo System's performance and ability to automatically detect wheezing in cough events, using the F1-score (value between 0 and 1), higher scores mean a better outcome.
Cough frequency (Matthews correlation coefficient)
24 hours
Measure C-mo System's performance and ability to automatically assess cough frequency, based on the average "number of coughs per hour", using Matthews correlation coefficient (MCC). The MCC value ranges from -1 to 1, and in this case it indicates the level of agreement between C-mo System's output and the result obtained from expert analysis (considered to be the gold standard in this study). A MCC value of -1 indicates total disagreement, 0 indicates that C-mo System's classification is no better than random guessing, and 1 represents a perfect classification (total agreement between C-mo System's output and the gold standard). Hence, higher scores mean a better outcome.
Cough frequency (Cohen's Kappa Index)
24 hours
Measure C-mo System's performance and ability to automatically assess cough frequency, based on the average "number of coughs per hour", using Cohen's Kappa coefficient (κ). The κ value indicates the level of agreement between C-mo System's output and the result obtained from expert analysis (considered to be the gold standard in this study). It ranges from -1 (worst possible performance) to 1 (best possible performance). Hence, higher scores mean a better outcome.
Cough type percentage (Matthews correlation coefficient)
24 hours
Measure C-mo System's performance and ability to automatically assess cough type, based on the percentage of each cough type, using Matthews correlation coefficient (MCC). The MCC value ranges from -1 to 1, and in this case it indicates the level of agreement between C-mo System's output and the result obtained from expert analysis (considered to be the gold standard in this study). A MCC value of -1 indicates total disagreement, 0 indicates that C-mo System's classification is no better than random guessing, and 1 represents a perfect classification (total agreement between C-mo System's output and the gold standard). Hence, higher scores mean a better outcome.
Cough type percentage (Cohen's Kappa Index)
24 hours
Measure C-mo System's performance and ability to automatically assess cough type, based on the percentage of each cough type, using Cohen's Kappa coefficient (κ). The κ value indicates the level of agreement between C-mo System's output and the result obtained from expert analysis (considered to be the gold standard in this study). It ranges from -1 (worst possible performance) to 1 (best possible performance). Hence, higher scores mean a better outcome.
Wheezing detection (Matthews correlation coefficient)
24 hours
Measure C-mo System's performance and ability to automatically assess wheeze in cough, based on the percentage of cough events in which wheezing was identified, using Matthews correlation coefficient (MCC). The MCC value ranges from -1 to 1, and in this case it indicates the level of agreement between C-mo System's output and the result obtained from expert analysis (considered to be the gold standard in this study). A MCC value of -1 indicates total disagreement, 0 indicates that C-mo System's classification is no better than random guessing, and 1 represents a perfect classification (total agreement between C-mo System's output and the gold standard). Hence, higher scores mean a better outcome.
Wheezing detection (Cohen's Kappa Index)
24 hours
Measure C-mo System's performance and ability to automatically assess wheeze in cough, based on the percentage of cough events in which wheezing was identified, using Cohen's Kappa coefficient (κ). The κ value indicates the level of agreement between C-mo System's output and the result obtained from expert analysis (considered to be the gold standard in this study). It ranges from -1 (worst possible performance) to 1 (best possible performance). Hence, higher scores mean a better outcome.