Fig 1.
Text corpus descriptive statistics.
A: Number of documents per year. B: Distribution of the number of words per document.
Fig 2.
Adopters were categorized as innovators, early adopters, early majority, late majority and laggards depending on their adoption time.
Fig 3.
A: Most relevant words. B: Monthly observations of the topic importance over time, smoothed using a 12-month exponentially weighted moving average.
Fig 4.
A: Most relevant words. B: Monthly observations of the topic importance over time, smoothed using a 12-month exponentially weighted moving average.
Fig 5.
A: Most relevant words. B: Monthly observations of the topic importance over time, smoothed using a 12-month exponentially weighted moving average.
Fig 6.
Paragraph vector topic model diffusion indicators obtained from topic modeling compared to Google Trends indices.
A: wikipedia topic. B: tablet topic. C: virtual reality topic. The PVTM measures are smoothed using the 12-month exponentially weighted moving average of the topic importance over time. We found statistical evidence that the unsmoothed PVTM topic importance measures Granger cause GTI for each of the presented topics.
Table 1.
VAR Granger causality/block exogeneity wald test.