Abstract: An electrocardiogram analysis apparatus includes a machine learning part that has a machine learning model realized by machine learning that uses training electrocardiogram data of a patient with paroxysmal arrhythmia during a non-paroxysmal period during which no episode of paroxysmal arrhythmia occurs; an input processing part that inputs electrocardiogram data of a person to be analyzed, which is a subject of analysis, into the machine learning model; and an output control part that outputs, to an information terminal, abnormality information which is to be output from the machine learning model and is about whether the person to be analyzed has paroxysmal arrhythmia.

Abstract: An electrocardiogram analyzing apparatus has: an acquiring section acquiring a plurality of electrocardiogram waveforms; a classifying section classifying the plurality of electrocardiogram waveforms into a plurality of groups on a basis of shape similarity; an accepting section accepting a reference position of a predetermined type of wave designated on a representative waveform corresponding to at least one electrocardiogram waveform belonging to a selection group selected from the plurality of groups; and an analyzing section identifying correspondence between a plurality of positions, along a time axis, on the representative waveform of a group to which a subject-of-analysis electrocardiogram waveform belongs, and a plurality of positions on the subject-of-analysis electrocardiogram waveform along a time axis, and decides, as a position of the predetermined type of wave included in the subject-of-analysis electrocardiogram waveform, a position on the subject-of-analysis electrocardiogram waveform correspo