Univariate Screening Help

Pre-modeling screening: per-feature Pearson, Spearman and ANOVA F statistics against a target with FDR correction. Produces a derived dataset of significant features.

Requires: One dataset with a numeric target column.

Form fields accepted by this method's POST request

analysis_type=univariate_screening, dataset_id, fdr_threshold, target_col

No dataset yet? Upload a file.
Each feature is tested for association with this target.
Parameters & Options
Features below this FDR in any test are kept in the derived screened dataset.
Optional: fill missing numeric values before the analysis runs. Leave as default to keep the method's own handling.
Back to Methods