An unknown signature is a representation of a text whose author has not been identified. It records writing-style measurements, such as average words per sentence and average sentence complexity, so the text can be compared with signatures from known authors. section 11.1
Use an unknown signature in a simple authorship-identification process:
- Calculate features from texts written by known authors to create known signatures. section 11.1
- Calculate the same features for the mystery text to create its unknown signature. section 11.1
- Compare the unknown signature with the known signatures and use the closest match as the authorship guess. section 11.1
A mystery novel
Suppose the known texts include works by Arthur Conan Doyle and Charles Dickens. A mystery novel receives an unknown signature, which is compared with both known signatures. The closest match suggests an author, but the result can be wrong if the writer is new, writes in a different style, or the selected features miss important characteristics. section 11.1
