Elasticsearch - 使用 Java 根据匹配分数进一步过滤结果

Question

We are currently limiting to 50 documents when we are searching in. But score for each document could be very different. eg score of 1st object = 5, score of 50th object = 0.0001. In this case we need to filter out objects with so low score.

What kind of statistical distribution/formula should we use upon querying to the elasticsearch? I am thinking of standard deviation but not so sure?

Answer 1

From what I understand you want to have a minimum_score to reduce results. Did you read about the min_score parameter?