def svd():
from util import reader
r = reader.tabSepReader("u.data")
from util import split
train, test1 = split.split(r.getR(), 1234567890)
from recommender import svd
W, H = svd.learnModel(r.getMaxUid(), r.getMaxIid(),
0.0002, train, 1000, 25, r.numberOfTransactions)
import numpy as np
np.savez_compressed(
"SVDModelFile", W=W, H=H)
# modelf = np.load("BPRMFModelFile" + ".npz")
# W = modelf['W']
# H = modelf['H']
# modelf.close
from util import test
t = test.MFtest(W, H, train)
test.hitrate(test1, t.getRec, 10)
def slopeone():
from util import reader
r = reader.tabSepReader("u.data")
from util import split
train, test1 = split.split(r.getR(), 12320894329854567890)
from recommender import slopeone
so = slopeone.slopeone(train)
# import cPickle
# output = open("slopeone.npz", "wb")
# cPickle.dump(so, output, -1)
# output.close()
# inputfile = open("slopeone.npz", "rb")
# so = cPickle.load(inputfile)
# inputfile.close()
from util import test
test.auc(test1, so.getRec, r)
def mf():
from util import reader
r = reader.tabSepReader("u.data")
from util import split
train, test1 = split.split(r.getR(), 1234567890)
from recommender import mf
W, H = mf.learnModel(r.getMaxUid(), r.getMaxIid(), 0.01, 0.01, 0.01,
0.1, train, 150, 3, r.numberOfTransactions,
mf.logLoss, mf.dLogLoss)
import numpy as np
# np.savez_compressed(
# "BPRMFModelFile", W=W, H=H)
# modelf = np.load("BPRMFModelFile" + ".npz")
# W = modelf['W']
# H = modelf['H']
# modelf.close
from util import test
t = test.MFtest(W, H, train)
test.hitrate(test1, t.getRec, 10)