Fig 1.
The degree distribution of users and objects in (a) Movielens, (b) Netflix and (c) RYM networks. (d), (e) and (f) are d(k) vs k in Movielens, Netflix and RYM networks, respectively.
For the blue curve, k denotes the degree of users and d(k) denotes the average degree of the neighboring objects of these users. For the red curve, k denotes the degree of objects and d(k) denotes the average degree of the neighboring users of these objects.
Fig 2.
The illustration of the H-Hybrid method.
Users and items are marked with circles and squares, respectively. Shaded circles indicate the target user for whom recommendation is done.
Fig 3.
The (a) Ranking score, (b) Precision, (c) personlization and (d) novelty of the H-Hybrid method in parameter space (λ1, λ2) in Netflix network.
The dashed line marks the region where RS is better than the RS value achievable with O-Hybrid method.
Fig 4.
The (a) Ranking score, (b) Precision, (c) personlization and (d) novelty of the O-Hybrid method as a function of λ in Netflix network.
The green lines mark the optimal λ* of the O-Hybrid method and the red lines mark the optimal results of the H-Hybrid method.
Fig 5.
vs λ1 in (a) Movielens, (b) Netflix and (c) RYM data.
The line corresponding to λ1 = λ2 is plotted to guide eyes. In (d)(e)(f), the minimum RS* is obtained for of the upper panels.
Table 1.
The results of all the metrics for different recommendation algorithms.
The entries corresponding to the best performance over all methods are emphasized in black.
Table 2.
The results of all the metrics for the O-Hybrid and H-Hybrid algorithms under the three-fold data division.
The entries corresponding to the best performance over all methods are emphasized in black.