You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi,
I am building an app with merlion backend and have trained an isolation forest model on multivariate data. Now i want to provide my users a visual representation of my anomlay score. With a marker separating scores before the threshold and scores after the threshold.
for this I create a PDF plot representation of anomaly scores provided by my Isolation Forest model ( model.get_anomaly_score(data) ) and now i want create marker represent threshold. But my threshold is in Zscore value and to create a marker i would have to convert it into absolute score value. so that i am able to plot.
Since the score are processed using a calibrator i am unable to convert z-score to absolute score value. A guidance on how can i obtain this conversion would be helpful.
I have already tried getting the mean and standard deviation from the score data and then trying to convert z score to an absolute value using : Z = (x-μ)/σ
i want to achieve something like this :
The text was updated successfully, but these errors were encountered:
Hi,
I am building an app with merlion backend and have trained an isolation forest model on multivariate data. Now i want to provide my users a visual representation of my anomlay score. With a marker separating scores before the threshold and scores after the threshold.
for this I create a PDF plot representation of anomaly scores provided by my Isolation Forest model (
model.get_anomaly_score(data)
) and now i want create marker represent threshold. But my threshold is in Zscore value and to create a marker i would have to convert it into absolute score value. so that i am able to plot.Since the score are processed using a calibrator i am unable to convert z-score to absolute score value. A guidance on how can i obtain this conversion would be helpful.
I have already tried getting the mean and standard deviation from the score data and then trying to convert z score to an absolute value using : Z = (x-μ)/σ
i want to achieve something like this :
The text was updated successfully, but these errors were encountered: