Menu
I have a pandas DataFrame where one column contains the following elements:
[2,2.5,3,2,2.6,10,10.3,10,10.1,10.3,10],
is there a python function that can detect the sudden change from 2.6 to 10 from that list? I have read a little bit and R can do that. Is there a similar function in python?
274
asked Jan 25 '16 12:01
Change point detection is similar to time series outlier detection but differs in important ways. Change point detection identifies time steps when one model changes to a new model (such as a change in the mean value), and outlier detection identifies time steps that deviate significantly from a single model.
To put it simple, a change point divides a time series into two segments where each segment has its own statistical characteristics (e.g., mean, variance, etc.). Thus, the change point is located where the underlying characteristics change abruptly.
Offline changepoint detection (CPD) algorithms are used for signal segmentation in an optimal way. Generally, these algorithms are based on the assumption that signal's changed statistical properties are known, and the appropriate models (metrics, cost functions) for changepoint detection are used.
What Is It? Change-point analysis is a method for identifying thresholds in relationships between two variables. More specifically, it is an analytical method that attempts to find a point along a distribution of values where the characteristics of the values before and after the point are different.
IIUC you could use pct_change for that to find differencies between neighbours and then compare with your limit (whatever it'll be):
s = pd.Series([2,2.5,3,2,2.6,10,10.3,10,10.1,10.3,10])
In [105]: s.pct_change()
Out[105]:
0 NaN
1 0.250000
2 0.200000
3 -0.333333
4 0.300000
5 2.846154
6 0.030000
7 -0.029126
8 0.010000
9 0.019802
10 -0.029126
dtype: float64
In [107]: s[s.pct_change() > 1]
Out[107]:
5 10
dtype: float64
In [117]: s[s.pct_change() > 1].index.tolist()
Out[117]: [5]
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With
