Skip to main content


Fig. 6 | Journal of Big Data

Fig. 6

From: Library adoption in public software repositories

Fig. 6

Feature ablation tests for predicting library adoptions with 95% confidence intervals. Results are ordered within feature categories from lowest to highest performance. Commit features (C) provide a large amount of information. The reduced feature set CPL has the highest performance on average, followed closely by the full feature set CUPLS

Back to article page