LSHash
LSHash copied to clipboard
projection type
The code is using np.random.randn() times input vector. In the LSH paper survey, we are using either (Gaussian Distribution * input + bias)/W or (Uniform Distribution * input). I was wondering if we should change the distribution to uniform in the code?
I have seen that Gaussian Random Projection is on way to implement the random projections. What is this LSH paper survey?