random tools in statistical theory
- w parameter initialize with standard distribution
import numpy as np
np.random.normal(0,1.0,(r,c)) # 0 : mean ; 1.0 : standard deviation ; r : height of matrix of w ;
np.random.randn(r,c)
np.random.standard_normal((r,c))
- b parameter initialize with 0
np.zeros((r,c))
扫描二维码关注公众号,回复:
880244 查看本文章
- np.random.uniform(0,1.0,10) # 0 : low ; 1.0 : high ; 10 : number of output
- np.random.binomial(1,0.5) # 二项式分布;输出像扔硬币结果:0 或者1