Question 4 — AB-900 Practice Exam 12
For each Microsoft Purview classification method, drag it to the appropriate description that best matches how it identifies or categorizes data.
Each item may be used once, more than once, or not at all.
| Option | Correct Match |
|---|---|
| Identifies sensitive data by matching values against a secure, organization-provided dataset | [Gap 1] |
| Uses predefined patterns, checksums, and keyword evidence to detect regulated data | [Gap 2] |
| Applies classification by learning document structure and content from examples | [Gap 3] |
| Allows users to choose how data is categorized | [Gap 4] |
- Manual Classification
- Retention Policies
- Sensitivity Label Policies
- Sensitive Information Types
- Trainable Classifiers
- Exact Data Match (EDM)
Correct answers
- Gap 1: Exact Data Match (EDM)
- Gap 2: Sensitive Information Types
- Gap 3: Trainable Classifiers
- Gap 4: Manual Classification
Explanations
Correct Matches
Sensitive Information Types – These use regex patterns, confidence scoring, keyword evidence, and validation techniques to identify regulated data such as credit card numbers or Social Security numbers.
Trainable Classifiers – These rely on machine learning models trained with example documents so that similar content can be identified even when keywords are not present.
Exact Data Match (EDM) – This method compares content against a hashed, secure dataset supplied by the organization, enabling extremely precise detection with minimal false positives.
Manual Classification – This allows users to explicitly select how data should be categorized when automated detection is not required or appropriate.
Unused Item(s)
Retention Policies – These control how long data is kept or deleted, not how it is classified.
Sensitivity Label Policies – These determine label availability and behavior, not data identification methods.