TY - GEN
T1 - A comparison of feature ranking and rank aggregation techniques in automatic sleep stage classification based on polysomnographic signals
AU - Najdi, Shirin
AU - Gharbali, Ali Abdollahi
AU - Fonseca, José Manuel
N1 - This work was partially funded by FCT Strategic Program UID/EEA/ 00066/203 of UNINOVA, CTS and INCENTIVO/EEI/UI0066/2014 of UNINOVA.
PY - 2016
Y1 - 2016
N2 - Sleep quality is one of the most important measures of healthy life, especially considering the huge number of sleep-related disorders. Identifying sleep stages using multi-channel recordings like polysomnographic (PSG) signals is an effective way of assessing sleep quality. However, manual sleep stage classification is time-consuming, tedious and highly subjective. To overcome this, automatic sleep classification was proposed, in which pre-processing, feature extraction and classification are the three main steps. Since the classification accuracy is deeply affected by the features selection, in this paper several feature selection methods as well as rank aggregation methods are compared. Feature selection methods are evaluated by three criteria: accuracy, stability and similarity. For classification two different classifiers (k-nearest neighbor and multilayer feedforward neural network) were utilized. Simulation results show that MRMR-MID achieves highest classification performance while Fisher method provides the most stable rankings.
AB - Sleep quality is one of the most important measures of healthy life, especially considering the huge number of sleep-related disorders. Identifying sleep stages using multi-channel recordings like polysomnographic (PSG) signals is an effective way of assessing sleep quality. However, manual sleep stage classification is time-consuming, tedious and highly subjective. To overcome this, automatic sleep classification was proposed, in which pre-processing, feature extraction and classification are the three main steps. Since the classification accuracy is deeply affected by the features selection, in this paper several feature selection methods as well as rank aggregation methods are compared. Feature selection methods are evaluated by three criteria: accuracy, stability and similarity. For classification two different classifiers (k-nearest neighbor and multilayer feedforward neural network) were utilized. Simulation results show that MRMR-MID achieves highest classification performance while Fisher method provides the most stable rankings.
KW - Biomedical signal processing
KW - Feature ranking
KW - Feature selection
KW - Polysomnography
KW - Rank aggregation
KW - Sleep stage classification
UR - http://www.scopus.com/inward/record.url?scp=84973863998&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-31744-1_21
DO - 10.1007/978-3-319-31744-1_21
M3 - Conference contribution
AN - SCOPUS:84973863998
SN - 978-3-319-31743-4
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 230
EP - 241
BT - Bioinformatics and Biomedical Engineering - 4th International Conference, IWBBIO 2016, Proceedings
A2 - Ortuno, F.
A2 - Rojas, I.
PB - Springer-Verlag
CY - Cham
T2 - 4th International Work-Conference on Bioinformatics and Biomedical Engineering, IWBBIO 2016
Y2 - 20 April 2016 through 22 April 2016
ER -