feat(ranker): fix the idf

jina-ai · Apr 13, 2020 · e6227c0 · e6227c0
1 parent e1c6b90
commit e6227c0
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/jina/executors/rankers/tfidf.py b/jina/executors/rankers/tfidf.py
@@ -178,7 +178,7 @@ def get_idf(match_idx):
         """
         q_df, q_id = TfIdfRanker._get_df(match_idx)
         total_df = np.sum(q_df)
-        return {idx: np.log10((total_df + 1.) / (df + 0.5)) for idx, df in zip(q_id, q_df)}
+        return {idx: np.log10((total_df + 1.) / (df + 0.5)) ** 2 for idx, df in zip(q_id, q_df)}
 
     def get_tf(self, match_idx, match_chunk_meta):
         """Get the tf dictionary for query chunks that matched a given doc.