There are two ways to create a instance of LocalOutlierFactor

1. clf= LocalOutlierFactor(n_neighbors=n_neighbors, contamination=contamination, novelty=False)

in this case we can use : clf.fit_predict(X)

2. clf = LocalOutlierFactor(n_neighbors=n_neighbors, contamination=contamination, novelty=True)

in this case we should use : clf.predict(X)

thanks