Column Name | Data Type | Description |
---|---|---|
cause | VARCHAR | Input attribute (column) being evaluated as a potential causal variable. |
effect | VARCHAR | Input attribute (column) being evaluated as a potential effect variable. |
library_size | INTEGER | Size of the library evaluated. |
correlation | DOUBLE PRECISION | For numerical cause variables: correlation between the values predicted by the effect attribute and the true value of the cause attribute. For categorical cause variables: NULL |
jaccard_index | DOUBLE PRECISION | For categorical cause variables: Jaccard similarity index between the values predicted by the effect attribute and the true value of the cause attribute. For numerical cause variables: NULL |
lower_bound | DOUBLE PRECISION | Lower bound of the 95% confidence interval of the prediction contained in the correlation or jaccard_index column (whichever is not NULL). |
upper_bound | DOUBLE PRECISION | Upper bound of the 95% confidence interval of the prediction contained in the correlation or jaccard_index column (whichever is not NULL). |
effect_size | DOUBLE PRECISION | For the smallest library size used with each cause/effect pair:
estimated effect size of increasing the library size from the smallest
value to the largest value. For other library sizes: NULL For a numerical effect: Cohen's q statistic. For a categorical effect: difference between the two similarity measures. An effect_size greater than approximately 0.25 indicates a causal relationship. |
effect_size_sd | DOUBLE PRECISION | Standard deviation of the effect size. |