RELATIVITY ANALYTICS SPECIALIST | Exam Questions &
Correct Answers with Rationales |2025/2026 | Instant download!!
1. In Relativity Analytics, what is the primary purpose of clustering?
a. To apply concept search queries
b. To group documents based on conceptual similarity
c. To remove duplicate documents
d. To automatically generate coding decisions
→ Correct Answer: b (correct answer)
Rationale: Clustering organizes documents into conceptually similar groups to help reviewers understand
dataset themes.
2. What data type does Relativity require before creating an Analytics index?
a. Extracted text
b. TIFF images
c. Audio transcripts only
d. Native files
→ Correct Answer: a (correct answer)
Rationale: Analytics indexes are built strictly from extracted text fields.
3. Which Analytics feature uses Latent Semantic Indexing (LSI)?
a. Structured analytics
b. Email threading
c. Concept search
d. Active learning
→ Correct Answer: c (correct answer)
Rationale: Concept searching relies on LSI to compare semantic meaning instead of literal keywords.
4. What is the ideal minimum document count for running clustering?
a. 100
b. 1,000
c. 5,000
d. 50
→ Correct Answer: b (correct answer)
Rationale: Clustering performs best with datasets containing roughly 1,000 or more documents.
5. What does the "coherence score" represent in clustering?
,a. Precision of coding predictions
b. Degree of semantic similarity within a cluster
c. Document review speed
d. OCR accuracy score
→ Correct Answer: b (correct answer)
Rationale: Coherence measures similarity of documents inside a cluster; higher means tighter conceptual
grouping.
6. Which operation identifies conversational relationships using email metadata?
a. Categorization
b. Email threading
c. Cluster visualization
d. Active learning
→ Correct Answer: b (correct answer)
Rationale: Email threading reconstructs email chains using metadata fields.
7. What is a “root email” in threading?
a. Any forwarded message
b. The last email in the thread
c. The earliest message in a conversation
d. An attachment to an email
→ Correct Answer: c (correct answer)
Rationale: The root email is the thread’s first message chronologically.
8. What field is required for categorization training?
a. Family ID
b. Coded yes/no field
c. Extracted text
d. Cluster ID
→ Correct Answer: c (correct answer)
Rationale: Categorization requires extracted text to build conceptual profiles.
9. What is the maximum number of categories allowed per categorization set?
a. 5
b. 25
c. 50
d. 200
→ Correct Answer: d (correct answer)
Rationale: Up to 200 categories may be created for a classification project.
10. What represents a good "training richness" value for categorization?
, a. 1–2%
b. 10–20%
c. 50–75%
d. 90%
→ Correct Answer: b (correct answer)
Rationale: Ideal richness for meaningful categorization is around 10–20%.
11. In an Analytics index, what is the function of the “Remove English stop words” option?
a. To delete metadata fields
b. To eliminate common words that reduce conceptual clarity
c. To remove email signatures
d. To filter out repetitive phrases
→ Correct Answer: b (correct answer)
Rationale: Stop words are removed to focus the index on meaningful text.
12. What is a recommended maximum number of documents for an Active Learning project?
a. 10,000
b. 500,000
c. 5 million
d. Unlimited
→ Correct Answer: b (correct answer)
Rationale: Active Learning can handle up to roughly 500k documents for optimal performance.
13. What Active Learning model type identifies documents likely to be relevant?
a. Categorical regression
b. Binary classification
c. Quantitative clustering
d. Logistic ranking
→ Correct Answer: b (correct answer)
Rationale: Active Learning uses a binary model (relevant/not relevant).
14. In email threading, what is a “duplicate spare”?
a. Missing metadata document
b. A document with no body text
c. A near-duplicate email with identical metadata
d. A duplicate beyond the primary copy in a thread
→ Correct Answer: d (correct answer)
Rationale: Duplicate spares are extra copies of a duplicated email beyond the canonical one.
15. Which of the following improves categorization training quality?
Correct Answers with Rationales |2025/2026 | Instant download!!
1. In Relativity Analytics, what is the primary purpose of clustering?
a. To apply concept search queries
b. To group documents based on conceptual similarity
c. To remove duplicate documents
d. To automatically generate coding decisions
→ Correct Answer: b (correct answer)
Rationale: Clustering organizes documents into conceptually similar groups to help reviewers understand
dataset themes.
2. What data type does Relativity require before creating an Analytics index?
a. Extracted text
b. TIFF images
c. Audio transcripts only
d. Native files
→ Correct Answer: a (correct answer)
Rationale: Analytics indexes are built strictly from extracted text fields.
3. Which Analytics feature uses Latent Semantic Indexing (LSI)?
a. Structured analytics
b. Email threading
c. Concept search
d. Active learning
→ Correct Answer: c (correct answer)
Rationale: Concept searching relies on LSI to compare semantic meaning instead of literal keywords.
4. What is the ideal minimum document count for running clustering?
a. 100
b. 1,000
c. 5,000
d. 50
→ Correct Answer: b (correct answer)
Rationale: Clustering performs best with datasets containing roughly 1,000 or more documents.
5. What does the "coherence score" represent in clustering?
,a. Precision of coding predictions
b. Degree of semantic similarity within a cluster
c. Document review speed
d. OCR accuracy score
→ Correct Answer: b (correct answer)
Rationale: Coherence measures similarity of documents inside a cluster; higher means tighter conceptual
grouping.
6. Which operation identifies conversational relationships using email metadata?
a. Categorization
b. Email threading
c. Cluster visualization
d. Active learning
→ Correct Answer: b (correct answer)
Rationale: Email threading reconstructs email chains using metadata fields.
7. What is a “root email” in threading?
a. Any forwarded message
b. The last email in the thread
c. The earliest message in a conversation
d. An attachment to an email
→ Correct Answer: c (correct answer)
Rationale: The root email is the thread’s first message chronologically.
8. What field is required for categorization training?
a. Family ID
b. Coded yes/no field
c. Extracted text
d. Cluster ID
→ Correct Answer: c (correct answer)
Rationale: Categorization requires extracted text to build conceptual profiles.
9. What is the maximum number of categories allowed per categorization set?
a. 5
b. 25
c. 50
d. 200
→ Correct Answer: d (correct answer)
Rationale: Up to 200 categories may be created for a classification project.
10. What represents a good "training richness" value for categorization?
, a. 1–2%
b. 10–20%
c. 50–75%
d. 90%
→ Correct Answer: b (correct answer)
Rationale: Ideal richness for meaningful categorization is around 10–20%.
11. In an Analytics index, what is the function of the “Remove English stop words” option?
a. To delete metadata fields
b. To eliminate common words that reduce conceptual clarity
c. To remove email signatures
d. To filter out repetitive phrases
→ Correct Answer: b (correct answer)
Rationale: Stop words are removed to focus the index on meaningful text.
12. What is a recommended maximum number of documents for an Active Learning project?
a. 10,000
b. 500,000
c. 5 million
d. Unlimited
→ Correct Answer: b (correct answer)
Rationale: Active Learning can handle up to roughly 500k documents for optimal performance.
13. What Active Learning model type identifies documents likely to be relevant?
a. Categorical regression
b. Binary classification
c. Quantitative clustering
d. Logistic ranking
→ Correct Answer: b (correct answer)
Rationale: Active Learning uses a binary model (relevant/not relevant).
14. In email threading, what is a “duplicate spare”?
a. Missing metadata document
b. A document with no body text
c. A near-duplicate email with identical metadata
d. A duplicate beyond the primary copy in a thread
→ Correct Answer: d (correct answer)
Rationale: Duplicate spares are extra copies of a duplicated email beyond the canonical one.
15. Which of the following improves categorization training quality?
