Microsoft Azure AI Fundamentals (AI-900)Describe features of Natural Language Processing (NLP) workloads on AzureMedium
A marketing team wants to analyze a large volume of customer feedback, including surveys and social media posts, to understand the predominant themes and recurrent issues without explicitly defining keywords or categories beforehand. They need an NLP technique that can automatically discover the underlying structure and main subjects discussed. Which NLP capability is best suited for this task?
- AText Classification
- BText Summarization
- CNamed Entity Recognition (NER)
- DTopic Modeling
Show answer & explanationAnswer & explanation
Correct answer: D. Topic Modeling
Topic Modeling is an unsupervised learning technique that can automatically discover abstract 'topics' (clusters of words that often co-occur) within a collection of documents, making it ideal for identifying predominant themes without pre-defined categories.
Why the other options are wrong
- A. Text Classification assigns documents to pre-defined categories, which is not suitable when categories are unknown.
- B. Text Summarization condenses documents, it doesn't discover underlying themes.
- C. NER identifies specific named entities (people, places), not abstract themes.
Topic Modeling
An unsupervised machine learning technique used to discover the abstract 'topics' that occur in a collection of documents. It identifies clusters of words that frequently appear together.
- Unsupervised: does not require pre-labeled data.
- Discovers latent themes or subjects.
- Useful for understanding large text corpora without prior knowledge of content.
Memory trick: Topics emerge from text like clouds from vapor.