Figure 1.
Two different ways to connect the same products, which are characterized by the capabilities needed to produce them.
On the left, we consider a hierarchical relationship; on the right, we join similar products.
Figure 2.
On the right, the resulting Taxonomy Network. One can notice how the presence of products in the export baskets of the countries influences their position in the network. For example, the ubiquitous product 2 becomes the root.
Figure 3.
The peculiar triangular shape of the matrix and the empirical distribution of products in the ubiquity-complexity plane are well reproduced by our model.
Table 1. Comparison among three different ways to reconstruct a taxonomy network.
Figure 4.
The largest component of the taxonomy network built from the 1995–2010 database.
The colors refer to the value of the matrix for Sweden, year 2010: green is 1, red is 0. The diameter of the vertices is proportional to the logarithm of the product complexity, as defined in [24]. Already from a visual inspection one could argue that a good strategic move for Sweden could be to produce the red, high complexity product in the Lab Equipment community.
Figure 5.
The disposition and the fitness for each country.
There is a clear correlation between the two variables, indicating a link between the growth potential of a country and its disposition on the taxonomy network.
Figure 6.
A component of the Taxonomy Network.
All nodes are clearly member of the same technological community, but umbrellas, whose presence is due to noise.
Figure 7.
An example of the time evolution of a component of the Taxonomy Network.
The studied country is South Korea. The red fillings represent an Increase of the RCA value. One can notice the diffusion from the center (root product) towards the borders of the component.
Figure 8.
The Taxonomy Network correctly finds the empirically most active edges, as empirically calculated with the Assist Matrix, using two different approaches which are described in the text.
Here we show the distribution of the rankings of the Assist Matrix elements selected by the Taxonomy Network. The rankings are calculated ordering the rows of the Assist Matrix from the largest element to the smallest one. The resulting distribution is peaked around small values, implying that a large fraction of the links suggested by our network correspond to the largest elements of the Assist Matrices.