Text analytics is the process of uncovering hidden patterns from raw human language to enable better decision-making and predictions.
This section goes through four real-world applications of text analytics:
Topic modeling
Classification
Sentiment analysis
Summarization
Topic Modeling
Identify topics that reflect underlying patterns and relationships from raw text data in a collection of documents.
Real-world applications:
Get insight into underlying causes of damages or failures and their correlations using maintenance and repair reports and work orders for manufacturing equipment, cars, or aircraft.
Identify major pain points or gaps for informing product or process design teams with customer surveys, reviews from social media, blogs, forums, and customer and technical support cases.
Classification
Classify documents into predetermined categories for efficient information retrieval and prediction.
Real-world applications:
Recommend actions or repair work needed based on text describing malfunctions or faults using a predictive model built with historical data.
Detect fraudulent reports using features extracted from text for cybersecurity or for asset selection using a prebuilt classification model.
Sentiment Analysis
Identify and score sentiments expressed in text.
Real-world applications:
Identify pain points and gaps for better product or process design using sentiment scores derived from customer surveys and social media.
Build an asset selection model for trading using sentiment scores of financial reports and news articles.
Summarization
Extract a summary from one or more documents automatically.
Real-world application:
Identify opportunities and gaps in scientific research by summarizing technical articles.
Highlight and understand relevant information faster by summarizing internal reports.
Test your knowledge!
Building a model that can predict recommended actions using text descriptions of the problems and actions taken in the past is an example of: