Implement knowledge mining and information extraction solutions →mediumMultiple Choice

AI-102 Practice Question: Implement knowledge mining and information extraction solutions

Exhibit

{
  "searchResults": {
    "value": [
      {
        "@search.score": 2.345,
        "content": "The quick brown fox jumps over the lazy dog.",
        "metadata_storage_path": "https://storage.blob.core.windows.net/documents/doc1.pdf"
      },
      {
        "@search.score": 1.234,
        "content": "A fast brown fox leaps over a sleepy dog.",
        "metadata_storage_path": "https://storage.blob.core.windows.net/documents/doc2.pdf"
      }
    ]
  }
}

Refer to the exhibit. You execute a search query on an Azure AI Search index and get these results. The query was 'brown fox'. Why is the first result scored higher than the second?

⚠ Common exam trap

AI-102 often tests the difference between default scoring and optional features like semantic ranking or vector search — candidates may assume advanced features are always active, but the default behavior is based on term frequency and proximity.

Answer choices

Why each option matters

Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.

Correct answer & explanation

✓

The first document has a higher term frequency and better term proximity for the query terms

The first document is scored higher because it has a higher term frequency and better term proximity for the query terms 'brown fox'. In Azure AI Search's default scoring profile, documents that contain the query terms more frequently and closer together receive higher relevance scores.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

✗
The first document has a higher value in a scoring profile field
Why it's wrong here
A scoring profile boosts only fields you explicitly weight, and the exhibit shows no such profile applied. Scoring profiles exist to promote documents by custom criteria, such as recency or a featured field, and would be the right answer if the query specified one. Here, term frequency drives the difference.
✗
The first document is more similar to the query in vector space
Why it's wrong here
The exhibit shows a keyword query scored by BM25 term frequency and inverse document frequency, not vector similarity; no vector field or embedding is queried. Vector search is tempting because it ranks by cosine distance between embeddings, and would be correct had the query used a vectorisable field with a configured vectoriser.
✗
The first document was boosted by a semantic ranking function
Why it's wrong here
Semantic ranking reranks the top results using a language model and returns semantic captions and answers, which the exhibit does not show; the ordering reflects lexical BM25 scoring. Semantic ranking is tempting because it does alter result order, and would be correct had semantic configuration been enabled on the query.
✓
The first document has a higher term frequency and better term proximity for the query terms
Why this is correct
Azure AI Search's BM25 scoring rewards documents where query terms appear more often and closer together. The first document contains 'brown' and 'fox' repeatedly and adjacent, so its term frequency and proximity boosts outweigh the second document's weaker matches.

Go deeper

AI-102 Practice Question: Implement knowledge mining and information extraction solutions

Exhibit

Why each option matters

Option-by-option breakdown

Related to this question