Although a lot effort has been devoted over the past years to the accurate measurement of echocardiographic deformation curves in order to quantify regional myocardial function, much less attention has been paid to the problem of dealing with missing or artifactual curves. Considering the difficulties associated with missing or unreliable curves in the clinical diagnostic process, this study sought to examine the usefulness of the K-nearest neighbor (KNN) imputation algorithm to address this problem. Experiments with segmental strain (rate) curves of 30 normal subjects showed that the imputation algorithm can lead to low estimation errors even with a high percentage of missing data.
Tabassian M., Alessandrini M., Jasaityte R., De Marchi L., Masetti G., D'Hooge J. (2016). Handling missing strain (rate) curves using K-nearest neighbor imputation. IEEE Computer Society [10.1109/ULTSYM.2016.7728809].
Handling missing strain (rate) curves using K-nearest neighbor imputation
Tabassian M.;Alessandrini M.;De Marchi L.;Masetti G.;
2016
Abstract
Although a lot effort has been devoted over the past years to the accurate measurement of echocardiographic deformation curves in order to quantify regional myocardial function, much less attention has been paid to the problem of dealing with missing or artifactual curves. Considering the difficulties associated with missing or unreliable curves in the clinical diagnostic process, this study sought to examine the usefulness of the K-nearest neighbor (KNN) imputation algorithm to address this problem. Experiments with segmental strain (rate) curves of 30 normal subjects showed that the imputation algorithm can lead to low estimation errors even with a high percentage of missing data.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
