def recommend(self, uids):
print 'Recommending'
bestMethod = self.best_alg()
alg = AlgFactory.create(bestMethod)
alg.fit(self.samples, self.targets)
recommendList = []
for u in uids:
recommendList.append(alg.recommend(u))
return recommendList
def recommendAll(self):
print 'Recommending'
# bestMethod = self.best_alg()
# alg = AlgFactory.create(bestMethod)
alg = AlgFactory.create('ItemCF')
uids = list(set(np.array(self.samples)[:,0]))
alg.fit(self.samples, self.targets)
recommendList = []
for u in uids:
rec = alg.recommend(u)
rec.insert(0, u)
recommendList.append(rec)
t = pd.DataFrame(recommendList)
t.to_csv('rec_list')
def mulProcess(self,result,processParameters):
algName = processParameters[0]
parameters = processParameters[1]
alg = AlgFactory.create(algName)
rec_cv = StratifiedKFold(self.labels, 2)
clf = grid_search.GridSearchCV(alg, parameters, cv=rec_cv)
clf.fit(self.samples, self.targets)
print(clf.best_estimator_)
print(clf.grid_scores_)
self.Lock.acquire()
result.append(algName)
result.append([clf.best_estimator_,clf.best_score_])
result.append(clf.grid_scores_)
self.Lock.release()
