Can anyone help me to calculate PALIVE(the probability that they are still alive at the end of the calibration period.) in Python?
I know R has a Pareto negative binomial distribution function pnbd.PAlive(params, r, s, apha, beta)
but what is similar to it in Python?
You could use conditional_probability_alive method from lifetimes package. You need to pass frequency, recency, and T for each customer. For example for BetaGeoFitter (BG/NBD model):
from lifetimes import BetaGeoFitter
from lifetimes.datasets import load_cdnow_summary
# load data
data = load_cdnow_summary(index_col=[0])
print(data.head())
# fit lifetimes model
bgf = BetaGeoFitter(penalizer_coef=0.0)
bgf.fit(data['frequency'], data['recency'], data['T'])
print(bgf)
# predict p_alives for customers
p_alive1 = bgf.conditional_probability_alive(2, 30.43, 38.86)
p_alive2 = bgf.conditional_probability_alive(1, 30, 30)
print(p_alive1, p_alive2)
Output:
frequency recency T
ID
1 2 30.43 38.86
2 1 1.71 38.86
3 0 0.00 38.86
4 0 0.00 38.86
5 0 0.00 38.86
<lifetimes.BetaGeoFitter: fitted with 2357 subjects, a: 0.79, alpha: 4.41, b: 2.43, r: 0.24>
0.7266084620654866 0.753658243186767