Case Folding
Reducing all letters to lower case (02 Pre-processing s.29).
- IR: fold to lower case because users tend to type lower case (Car, CAR -> car).
- Possible exceptions (mid-sentence upper case): General Motors vs general motors; Fed (Federal Reserve) vs fed; SAIL (Stanford Artificial Intelligence Language, etc) vs sail.
- Sentiment Analysis and Information Extraction: case is helpful. US vs us: “US won a gold medal”; “They like US.” vs “They like us.”
Related
Concept - Text Normalization | Concept - Sentiment Analysis | Concept - Information Extraction