Matrix Factorization & Collaborative Filtering

Contents

Matrix Factorization & Collaborative Filtering#

Classic and neural collaborative filtering models that learn user/item representations from interactions.

Production models (CI-tested)#

Model	Type	Import	Tutorial
SAR	Item similarity / co-occurrence	`from corerec.engines.collaborative import SAR`	SAR

SAR (DataFrame API)#

SAR expects Microsoft Recommenders-style column names: userID, itemID, rating.

from corerec.engines.collaborative import SAR
import pandas as pd

df = pd.DataFrame({
    "userID": [0, 0, 1, 1],
    "itemID": [10, 11, 10, 12],
    "rating": [5, 4, 3, 5],
})
model = SAR(similarity_type="jaccard")
model.fit(df)
recs = model.recommend(user_id=0, top_k=10)

NCF (DataFrame API)#

from corerec.engines.collaborative import NCF

model = NCF(num_epochs=10, verbose=True)
model.fit(df)  # columns: user_id, item_id, rating
recs = model.recommend(user_id=0, top_k=10)

FAST / FASTRecommender (triplet API)#

from corerec.engines.collaborative import FAST

model = FAST(factors=64, iterations=15, seed=42)
model.fit(user_ids, item_ids, ratings)
score = model.predict(user_id=0, item_id=10)

Sandbox models (experimental)#

Model	Import	Tutorial
ALS	`corerec.sandbox.collaborative_full.mf_base.als_base`	ALS
SVD	`corerec.sandbox.collaborative_full.mf_base.svd_base`	SVD
A2SVD	`corerec.sandbox.collaborative_full.mf_base.a2svd_base`	A2SVD
MF-Base	`corerec.sandbox.collaborative_full.mf_base.matrix_factorization_base`	MF Base
FM-Base	`corerec.sandbox.collaborative_full.nn_base.FM_base`	FM Base
Matrix Factorization	`corerec.sandbox.collaborative_full.mf_base.matrix_factorization_base`	Matrix Factorization
User-Based CF	`corerec.sandbox.collaborative_full.mf_base.user_based_base`	User-Based

Legacy paths

Old docs referenced corerec.engines.unionizedFilterEngine.* — that module layout no longer exists. Use production imports above or sandbox paths from tutorials.

When to use#

Scenario	Start with
Fast baseline, implicit feedback	SAR
Neural pairwise ranking	NCF
Low-latency similarity CF	FAST
Research / classic MF	Sandbox ALS, SVD

See also#