Reputation: 97
Here is the way I could do using sklearn minmax_scale, however sklearn can not be able to integrate with pyspark. Is there anyway, I could use an alternate way in spark for minmax scaling on an array? Thanks.
for i, a in enumerate(np.array_split(target, count)):
start = q_l[i]
if i == (count - 1):
end = 1.0
else:
end = q_l[i + 1]
target_scaled = minmax_scale(a, feature_range=(start, end))
result.append(target_scaled)
results = np.concatenate(results)
Upvotes: 1
Views: 219