Referring now to FIG. 4, an example graph 400 of viewing patterns generated from viewing history data as well as video content similarity is shown. To generate the graph 400, multiple viewing histories were compiled and viewings, V, were paired together as described above with reference to FIG. 3. To generate the data points for graph 400, however, a similarity score, S, was calculated for the video content, C_x, and C_y, of each pairing. Determining similarity scores is discussed in further detail below. The similarity score, S, was thus included with the video content, C_x, and C_y, and the time difference, ΔT, between viewings to obtain an extended pairing, P′. Accordingly the extended pairing, P′, may be designated using the 4-tuple, P′={C_x, C_y, ΔT, S}.

Each data point in the graph 400 thus corresponds to an extended pairing, P′. The x-axis 402 plots each extended pairing, P′, as the number of hours elapsed, ΔT, between viewings of the paired video content, C_x, and C_y. The y-axis 404 plots the frequency of the similarity score, S, in the extended pairings generated. The graph 400 thus illustrates the total number of extended pairings, P′, where the video content, C_x, and C_y, have a similarity score, S, and were viewed ΔT hours apart.