Is it possible to turn the following Python code loop into code without loops in Python?
steps = 10000
rho = 0.9
e = np.random.randn(steps)
y = [0]
# Possible to get rid of this loop?
for i in range(len(e)):
y.append(rho*y[i] + e[i])
The code is to generate simple autoregressive (order 1) time series where given some starting value, every next value is rho time previous plus some random noise.
Check out this code :
import numpy as np
steps = 10000
rho = 0.9
np.random.seed(0)
e = np.random.randn(steps)
z = list(np.zeros(steps))
z[:] = map(lambda x, y: rho*z[y] + x, list(e), list(range(steps)))
Method-2 :
m = list(np.zeros(steps))
m[:] = map(lambda x, y: m.append(rho*m[y] + x), list(e), list(range(steps)))
m = m[10000:]
print(m)
Some quick benchmarks from ipython
In [1]: timeit z[:] = map(lambda x, y: rho*z[y] + x, list(e), list(range(steps)))
1.27 ms ± 19.3 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
In [2]: timeit m[:] = map(lambda x, y: m.append(rho*m[y] + x), list(e), list(range(steps)))
1.39 ms ± 31.9 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
In [3]: timeit for i in range(len(e)): y.append(rho*y[i] + e[i])
8.18 ms ± 270 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)