1. Keyword Extraction (Key Phrase Extraction)
📌 What it does
Identifies the most important words or phrases in a piece of text.
🎯 Purpose
Quickly understand the main topics
Tag or categorise documents
✅ Example
Text: “Microsoft Fabric enables scalable data analytics and integration”
➡️ Extracted keywords:
Microsoft Fabric
data analytics
integration
🧩 When to use
Search indexing
Document tagging
Topic identification
🧾 2. Entity Detection (Named Entity Recognition – NER)
📌 What it does
Finds and classifies real-world items in text.
🏷️ Common entity types
People
Organisations
Locations
Dates
Products
✅ Example
“Microsoft announced a new product in London on 5 May 2026”
➡️ Entities:
Organisation: Microsoft
Location: London
Date: 5 May 2026
🧩 When to use
Extracting structured data
Building knowledge graphs
Compliance / document processing
😊 3. Sentiment Analysis
📌 What it does
Determines the opinion or emotion expressed in text.
🎯 Output
Positive
Negative
Neutral
(Sometimes a confidence score)
✅ Example
“The service was excellent and very fast”
➡️ Sentiment: Positive
🧩 When to use
Customer feedback analysis
Social media monitoring
Product reviews
✂️ 4. Summarization
📌 What it does
Creates a shortened version of text while preserving key meaning.
🧠 Types (high-level)
Extractive → pulls key sentences
Abstractive → generates new summary text (more advanced)
✅ Example
Long report → Short paragraph summary of key points
🧩 When to use
News summaries
Meeting notes
Document previews
⚖️ 5. How to Identify the Right Technique (Exam Logic)
Scenario Technique
“Find important topics in documents” Keyword extraction
“Extract names, places, dates” Entity detection
“Understand opinion or emotion” Sentiment analysis
“Shorten long content” Summarization
⚠️ 6. Common Exam Traps
Keyword extraction vs Entity detection
Keywords = important phrases
Entities = specific real-world items
Sentiment vs Summarization
Sentiment = opinion
Summarization = shorter version
Entity detection vs Information extraction
Entity detection = identify items
Info extraction = structured outputs (broader)
🧠 7. Simple Memory Trick
“Key → Who → Feel → Short”
Key → Keyword extraction
Who → Entity detection
Feel → Sentiment
Short → Summarization
✅ Summary
To describe common text analysis techniques in AI-901:
Keyword extraction → identifies important topics
Entity detection → finds people, places, dates, etc.
Sentiment analysis → determines opinion
Summarization → shortens text