It is an object of the present invention to enable prediction of a menstrual date while preventing a decrease in the accuracy of the prediction, even if there is a large variation among menstrual cycles. An information processing device determines a threshold vale used to predict a menstrual cycle, based on the result of a comparison between a base cycle and a representative value of a plurality of reference cycles modified based on each of a plurality of candidate threshold values for an outlier. Alternatively, the information processing device removes a first outlier identified based on a first threshold value from among a plurality of past menstrual cycles, changes a second outlier based on a second threshold value to a length corresponding to a second threshold value, and then predicts the length of a menstrual cycle.
