Figure 1.
Number of actors per year and number of movies per year for the entire IMDB database (lines) and for the adult IMDB database (symbols).
In the plot the adult IMDB points have been multiplied by a factor of 5 for clarity. Inset: the same numbers on a logarithmic y-scale.
Figure 2.
Comparison of the IMDB network to the adult IMDB network.
a) Probability distribution that a number of actors participated in a movie (as reported in IMDB). b) Distribution of the links weights, or equivalently, the probability that two actors have participated in
movies together.
Figure 3.
Snapshots of the ‘adult IMDB’ network, where two actors are connected if they have co-starred in at least a) , b)
, c)
, or d)
movies, respectively.
Only the largest connected component is shown, and the corresponding network sizes are , 14444, 5315, and 2100 actors, respectively.
Figure 4.
Degree distribution of the adult IMDB network, for different connection strengths, , 2, 4, or 8 movies, respectively.
The solid line has a slope, the exponent of the power law degree distribution, (see text).
Figure 5.
Probability distribution of the shortest paths in networks with varying weight threshold .
Inset: Maximum diameter in the network as a function of .
Figure 6.
Scaling of the number of boxes as a function of the maximum box diameter
, for the adult IMDB network and for different
values.
Except for , the other networks have a fractal dimension, shown by the dashed line with slope,
.
Figure 7.
Scaling of the modularity as a function of the maximum box diameter
, for the adult IMDB network and for different
values.
All networks scale with , with a slope close to
, as shown by the dashed line with the same slope value.
Figure 8.
Networks resulting from the box-covering procedure at different values.
Top row: Adult IMDB network at . Bottom row: Adult IMDB network at
.
Figure 9.
Variation of the average degree in the IMDB network for different values, as a function of the average box mass
.