Machine learning uses algorithms that learn patterns from structured data to predict outcomes without being explicitly programmed for each pattern. In text mining, natural language processing first converts language into forms a computer can analyze; machine learning then helps identify patterns, classify content, or predict sentiment. section 5.1
Machine learning supports practical text-analysis tasks such as:
- Analyzing sentiment more effectively than frequency counts, keyword searches, or dictionary-based methods for many everyday tasks. section 6.2.4
- Finding patterns and predicting outcomes from text prepared through natural language processing. section 5.1
- Working across social media, websites, emails, company documents, and academic publications as sources of unstructured text. section 5.1
Sentiment analysis
A sentiment-analysis task can use a machine-learning model to recognize patterns in text and classify its emotional orientation. Unlike a simple keyword or dictionary approach, the model learns from structured examples. Pretrained models make this practical without requiring the analyst to construct the model or master its underlying mathematics. section 6.2.4
