aayushkumarjvs/RoroDataInternship
In univariate, random samples with step function the after plot there exists amount of discords
As the plot contains of 10,000 samples of random samples of data with step functions that generate unequal patters in between with a range of (0,1]
Most of the anomailes are detected from a prior existing piece of code that is the z-normalized script and the discord script, which consists of two functions which are as follows:-
Of these two we even have the flexibility of changing our parameters which are the selection of window size, z_threshold, length of the series
find_discords_brute_force(series, win_size, num_discords=2, z_threshold=0.01):
find_best_discord_brute_force(series, win_size, global_registry, z_threshold=0.01):
