Alternative implementation of the robust peak detection algorithm by van Brakel 2014 Classifies data points with modified z-scores greater than max_score as outliers ccording to Iglewicz and Hoaglin 1993
detect_peak_modz(df_comparison, lag = 5, max_score = 3.5)
output of compare_flags
width of interval to compute rolling median and MAD, defaults to 5
modified z-score cutoff to classify observations as outliers, defaults to 3.5
df with classifications