The following question was deleted due to some technical issues. Here you may find the question and our answer.
The challenge data gives the patient labels, but also the location of the patient labels on each record. But there is such a situation, assuming that a patient has four records, and the label of the patient is present, but only the position of the label on one of the records is given in the .txt file, such as '"AV", not given The information of the other three records is displayed. At this time, can we understand that the labels of the other three records are absent.
There are cases where the patient has murmur, but the human annotator could not hear the murmur in every location. We believe you are talking about these kinds of cases. For example, there are four available locations for some cases, but the human annotator could hear the murmur only in 2 of them. In these cases, one can label two of the recordings as "present" and the other two as "absent". However, at the end, the Challenge considers the label of the patient, not the label of each recording or location. You can train your machine learning algorithms as you like, but the cost of the Challenge is based on the patient-level diagnosis.
(On behalf of the Challenge team.)
Please post questions and comments in the forum. However, if your question reveals information about your entry, then please email Challenge at physionet.org. We may post parts of our reply publicly if we feel that all Challengers should benefit from it. We will not answer emails about the Challenge to any other address. This email is maintained by a group. Please do not email us individually.