Calculate Uzzi novelty score

Computes the Uzzi et al. (2013) novelty score for each work based on atypical combinations of cited journals. A work is novel if its reference list contains unusual journal pairings – combinations that are rarely seen together in the broader literature.

Usage

sm_metric_novelty(corpus, call = rlang::caller_env())

Arguments

corpus

An sm_corpus object with populated references and works tables. Works and their references should have source_id assignments.

call

Caller environment for error reporting.

Value

A tibble with columns work_id and novelty. Higher values indicate more novel (atypical) journal combinations. Works with insufficient data receive NA.

Details

The algorithm:

1. For each work, identify the journals (sources) of its cited references.

2. Enumerate all pairwise journal combinations in each reference list.

3. Compute the observed frequency of each journal pair across all works.

4. Compute the expected frequency under independence (product of marginal frequencies).

5. The novelty score for a work is the median of \(-\log_{10}(observed / expected)\) across all its journal pairs. High values indicate atypical combinations.

Works with fewer than 2 references with known journals receive NA.

References

Uzzi, B., Mukherjee, S., Stringer, M., & Jones, B. (2013). Atypical Combinations and Scientific Impact. Science, 342(6157), 468–472. doi:10.1126/science.1240474

See also

Other metrics: sm_metric_collab_index(), sm_metric_disruption(), sm_metric_fnci(), sm_metric_g_index(), sm_metric_h_index(), sm_metric_m_index(), sm_metric_rcr(), sm_self_citation(), sm_summary_authors(), sm_summary_period(), sm_summary_sources(), sm_summary_works()

Examples

corpus <- sm_example_corpus()
nov <- sm_metric_novelty(corpus)
head(nov)
#> # A tibble: 6 × 2
#>   work_id    novelty
#>   <chr>        <dbl>
#> 1 W000000001  -0.323
#> 2 W000000002  -0.384
#> 3 W000000003  -0.335
#> 4 W000000004  -0.313
#> 5 W000000005  -0.345
#> 6 W000000006  -0.378