TF-IDF and BM25 scoring in Elasticsearch - Step-by-Step Execution

Concept Flow - TF-IDF and BM25 scoring

Input Query

↓

Document Collection

↓

Calculate Term Frequency (TF)

↓

Calculate Inverse Document Frequency (IDF)

↓

Compute TF-IDF Score

↓

Apply BM25 Formula (TF, IDF, doc length)

↓

Rank Documents by Score

↓

Return Top Results

The flow shows how Elasticsearch scores documents by first calculating term frequency and inverse document frequency, then combining them using TF-IDF or BM25 formulas to rank documents.

Execution Sample

Elasticsearch

GET /my_index/_search
{
  "query": {
    "match": { "text": "apple banana" }
  }
}

This query searches documents containing 'apple' and 'banana' and scores them using BM25 by default.

Execution Table

Step	Action	Term	TF	IDF	BM25 Score	Explanation
1	Calculate TF for 'apple' in Doc1	apple	3	2.0	0	Count how many times 'apple' appears in Doc1
2	Calculate TF for 'banana' in Doc1	banana	1	1.5	0	Count how many times 'banana' appears in Doc1
3	Calculate IDF for 'apple'	apple	-	2.0	0	Inverse document frequency for 'apple'
4	Calculate IDF for 'banana'	banana	-	1.5	0	Inverse document frequency for 'banana'
5	Compute BM25 score for 'apple' in Doc1	apple	3	2.0	2.5	Apply BM25 formula with TF=3, IDF=2.0, doc length normalization
6	Compute BM25 score for 'banana' in Doc1	banana	1	1.5	1.2	Apply BM25 formula with TF=1, IDF=1.5, doc length normalization
7	Sum BM25 scores for Doc1	-	-	-	3.7	Total BM25 score for Doc1 is sum of term scores
8	Repeat steps 1-7 for Doc2	-	-	-	2.1	Calculate scores for another document
9	Rank documents by BM25 score	-	-	-	-	Doc1 (3.7) ranks higher than Doc2 (2.1)
10	Return top ranked documents	-	-	-	-	Results returned to user sorted by score

💡 All documents scored and ranked; top results returned.

Variable Tracker

Variable	Start	After Step 1	After Step 5	After Step 7	Final
TF_apple_Doc1	0	3	3	3	3
TF_banana_Doc1	0	1	1	1	1
IDF_apple	-	-	2.0	2.0	2.0
IDF_banana	-	-	1.5	1.5	1.5
BM25_apple_Doc1	0	0	2.5	2.5	2.5
BM25_banana_Doc1	0	0	0	1.2	1.2
BM25_total_Doc1	0	0	0	3.7	3.7

Key Moments - 3 Insights

Why does BM25 score consider document length while TF-IDF does not?

BM25 adjusts term frequency by document length to avoid bias toward longer documents, as shown in steps 5 and 6 where BM25 formula includes normalization, unlike simple TF-IDF.

Why is IDF important in scoring?

IDF reduces the weight of common terms across documents, making rare terms more influential, as seen in steps 3 and 4 where IDF values differ per term.

How are scores combined for multiple terms?

Scores for each term are calculated separately then summed to get the document's total score, demonstrated in step 7.

Visual Quiz - 3 Questions

Look at the execution table, what is the BM25 score for 'apple' in Doc1 at step 5?

A 1.5

B 3

C 2.5

D 0

💡 Hint

Check the BM25 Score column at step 5 in the execution_table.

At which step does the total BM25 score for Doc1 get calculated?

A Step 6

B Step 7

C Step 9

D Step 3

💡 Hint

Look for the row where BM25_total_Doc1 is updated in the variable_tracker and execution_table.

If the term frequency of 'banana' in Doc1 increased to 4, how would the BM25 score at step 6 change?

A It would increase

B It would decrease

C It would stay the same

D It would become zero

💡 Hint

Higher term frequency increases BM25 score as seen in steps 1, 2, 5, and 6.

Concept Snapshot

TF-IDF and BM25 scoring in Elasticsearch:
- TF counts term appearances in a doc
- IDF measures term rarity across docs
- TF-IDF multiplies TF by IDF
- BM25 improves TF-IDF by normalizing for doc length
- Elasticsearch uses BM25 by default to rank search results

Full Transcript

This visual execution shows how Elasticsearch scores documents using TF-IDF and BM25. First, it counts how often each search term appears in a document (TF). Then, it calculates how rare each term is across all documents (IDF). BM25 scoring combines these with adjustments for document length to avoid favoring longer documents. Each term's BM25 score is computed and summed to get the document's total score. Documents are then ranked by these scores to return the most relevant results. Key points include the importance of IDF to reduce common term weight, BM25's length normalization, and summing scores for multiple terms. The execution table traces these calculations step-by-step for clarity.

Practice

1. What is the main difference between TF-IDF and BM25 scoring in Elasticsearch?

easy

A. BM25 ignores inverse document frequency, TF-IDF uses it heavily.

B. TF-IDF is the default scoring algorithm in Elasticsearch, BM25 is optional.

C. BM25 considers term frequency saturation and document length normalization, TF-IDF does not.

D. TF-IDF scores are always higher than BM25 scores for the same query.

Solution

1. Step 1: Understand TF-IDF basics

   TF-IDF scores words by multiplying term frequency (TF) with inverse document frequency (IDF), but it does not adjust for document length or term frequency saturation.

2. Step 2: Understand BM25 improvements

   BM25 improves TF-IDF by adding term frequency saturation (diminishing returns for repeated terms) and normalizing by document length, making scores more balanced.

3. Final Answer:

   BM25 considers term frequency saturation and document length normalization, TF-IDF does not. -> Option C

4. Quick Check:

   BM25 adds saturation and length normalization [OK]

Hint: BM25 adjusts TF-IDF with saturation and length normalization [OK]

Common Mistakes:

• Thinking TF-IDF is default in Elasticsearch
• Believing BM25 ignores IDF
• Assuming TF-IDF always scores higher

2. Which Elasticsearch query parameter correctly switches the similarity scoring to TF-IDF instead of BM25?

easy

A. ""{
"similarity": "classic"
}"""

B. ""{
"similarity": "bm25"
}"""

C. ""{
"similarity": "tfidf"
}"""

D. ""{
"similarity": "default"
}"""

Solution

1. Step 1: Recall Elasticsearch similarity options

   Elasticsearch uses "bm25" by default. To switch to TF-IDF, you use the "classic" similarity setting.

2. Step 2: Match parameter to TF-IDF

   The "classic" similarity corresponds to TF-IDF scoring, so setting "similarity": "classic" enables TF-IDF.

3. Final Answer:

   { "similarity": "classic" } -> Option A

4. Quick Check:

   TF-IDF similarity is "classic" [OK]

Hint: Use similarity: classic to enable TF-IDF scoring [OK]

Common Mistakes:

• Using "tfidf" as similarity name (not valid)
• Assuming "default" switches to TF-IDF
• Confusing bm25 with TF-IDF

3. Given the following Elasticsearch query using BM25, what will happen to the score if a term appears multiple times in a document?

medium

A. The score increases but with diminishing returns due to term frequency saturation.

B. The score increases linearly with each additional term occurrence.

C. The score stays the same regardless of term frequency.

D. The score decreases as term frequency increases.

Solution

1. Step 1: Understand BM25 term frequency effect

   BM25 applies a saturation function to term frequency, so each additional occurrence adds less to the score than the previous one.

2. Step 2: Compare to linear increase

   Unlike TF-IDF which can increase linearly, BM25 limits the impact of repeated terms to avoid overvaluing them.

3. Final Answer:

   The score increases but with diminishing returns due to term frequency saturation. -> Option A

4. Quick Check:

   BM25 uses saturation, not linear increase [OK]

Hint: BM25 scores repeated terms with diminishing returns [OK]

Common Mistakes:

• Assuming linear score increase with term frequency
• Thinking score stays constant with repeats
• Believing score decreases with more terms

4. You configured Elasticsearch to use TF-IDF similarity but notice search results are not ranking as expected. Which of the following is a likely cause?

medium

A. TF-IDF requires a special query syntax different from standard queries.

B. TF-IDF does not use inverse document frequency, so scores are random.

C. BM25 must be disabled in the cluster settings to use TF-IDF.

D. You forgot to reindex the data after changing similarity settings.

Solution

1. Step 1: Understand similarity changes require reindexing

   Changing similarity settings affects how documents are scored, but Elasticsearch applies these only after reindexing the data.

2. Step 2: Identify common mistake

   If you don't reindex, old scores remain, so results won't reflect the new TF-IDF similarity.

3. Final Answer:

   You forgot to reindex the data after changing similarity settings. -> Option D

4. Quick Check:

   Reindexing needed after similarity change [OK]

Hint: Always reindex after changing similarity settings [OK]

Common Mistakes:

• Believing TF-IDF ignores IDF
• Thinking BM25 must be disabled cluster-wide
• Assuming query syntax changes for TF-IDF

5. You want to improve search relevance for a collection of very short documents using Elasticsearch. Which scoring approach and adjustment is best to handle short documents effectively?

hard

A. Use BM25 with a higher k1 parameter to ignore term frequency saturation.

B. Use BM25 with a lower b parameter to reduce length normalization effect.

C. Use TF-IDF similarity because it handles short documents better by default.

D. Use TF-IDF and disable inverse document frequency to boost short docs.

Solution

1. Step 1: Understand BM25 parameters for short docs

   BM25's b parameter controls length normalization. Lowering b reduces penalty on short documents, improving their scores.

2. Step 2: Compare to TF-IDF and other options

   TF-IDF lacks length normalization control. Increasing k1 ignores saturation, which can hurt relevance. Disabling IDF is incorrect.

3. Final Answer:

   Use BM25 with a lower b parameter to reduce length normalization effect. -> Option B

4. Quick Check:

   Lower b helps short docs in BM25 [OK]

Hint: Lower BM25 b parameter to favor short documents [OK]

Common Mistakes:

• Assuming TF-IDF is better for short docs by default
• Increasing k1 to ignore saturation hurts relevance
• Disabling IDF reduces search quality