Word Sequential Using Deep LSTM and Matrix Factorization to Handle Rating Sparse Data for E-Commerce Recommender System

Table 6

Performance comparison of LSTM-PMF with the state-of-the-art methods on ML-1M.

Sparseness level (high-low)	RMSE evaluation result			Comparison result
	PMF	CNN-PMF	LSTM-PMF	PMF versus LSTM-PMF (%)	CNN-PMF versus LSTM-PMF (%)
10% (90% sparseness level)	1.64697	0.99541	0.9928	39.00	0.26
20% (80% sparseness level)	1.26577	0.9276	0.93214	26.70	0.48
30% (70% sparseness level)	1.1118	0.90507	0.89993	18.59	0.56
40% (60% sparseness level)	1.03992	0.88525	0.88458	14.87	0.08
50% (50% sparseness level)	0.99064	0.87787	0.87114	11.38	0.76
60% (40% sparseness level)	0.95897	0.86774	0.86157	9.50	0.71
70% (30% sparseness level)	0.93369	0.86874	0.85471	6.95	1.61
80% (20% sparseness level)	0.91134	0.85574	0.84745	6.10	0.96
90% (10% sparseness level)	0.90452	0.84971	0.84079	6.06	1.04
∑ (total)				139	6.46
(average)				15.4	0.71