A financial institution wants to screen large volumes of customer emails for potential fraud indicators. The solution needs to identify specific phrases (e.g., 'urgent transfer', 'account compromised'), unusual tone, and report these findings with a confidence score. While Azure AI Language offers pre-built models, the institution has unique fraud terminology. Which approach should be used to achieve highly accurate detection of these specialized phrases and tones?
- AUtilize Azure AI Language's custom text classification and custom named entity recognition features.
- BDevelop a custom text classification model using Azure Machine Learning with a large dataset of fraud emails.
- CIntegrate Azure AI Search with keyword highlighting for fraud terms.
- DUse Azure AI Language's pre-built sentiment analysis and key phrase extraction.
Show answer & explanationAnswer & explanation
Correct answer: A. Utilize Azure AI Language's custom text classification and custom named entity recognition features.
Azure AI Language's custom text classification allows training a model to categorize text into custom labels (e.g., 'fraudulent', 'suspicious'). Custom Named Entity Recognition (NER) enables training a model to identify specific, domain-specific entities or phrases (like unique fraud terminology) that pre-built models might miss. This combination directly addresses the need for specialized detection and classification.
Why the other options are wrong
- B. While possible, developing a custom model from scratch in Azure ML is more complex and time-consuming than leveraging the focused capabilities of Azure AI Language for text classification and NER, particularly when starting with text data.
- C. Azure AI Search is for indexing and searching, not for advanced NLP tasks like custom classification or identifying nuances in tone and specific phrases beyond simple keyword matching.
- D. Pre-built models might not be accurate enough for highly specialized fraud terminology and complex tonal analysis beyond basic sentiment.
Azure AI Language Customization
The ability to train Azure AI Language models with domain-specific data to improve accuracy for custom text classification, named entity recognition, and conversational language understanding tasks.
- Enhances accuracy for niche terminology.
- Supports custom text classification and NER.
- Requires labeled data for training.
Memory trick: Custom Language AI finds the hidden fraud clues.