Your question is Detect Anomalies in Data. Start with the requirements on the right.
Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.
Trend Micro Vision One can surface unusual telemetry measurements such as process counts or network events. Given a sequence of numeric observations, identify values that are statistically anomalous using the median absolute deviation, or MAD.
Implement detect_anomalies(values, threshold). values is a non-empty list of integers or floating-point numbers, and threshold is a positive number. Return a list of the original zero-based indices whose values are anomalous, in ascending index order.
Calculate the median of values, then calculate the absolute deviation of every value from that median. The MAD is the median of those deviations. A value is anomalous when:
abs(value - median) > threshold * MAD
If the MAD is zero, return every index whose value differs from the median. Do not modify the input list.
def detect_anomalies(values, threshold):