Mis on ROUGE ja kuidas erineb ROUGE-L ROUGE-N-st?

ROUGE (Recall-Oriented Understudy for Gisting Evaluation) is a set of metrics for evaluating the quality of automatic summarization and text generation, comparing the generated text with a reference.

• ROUGE-N measures the matches of n-grams (sequences of n words) between the generated and reference texts. For example, ROUGE-1 for matches of individual words, ROUGE-2 for pairs of words.

• ROUGE-L is based on the length of the longest common subsequence (Longest Common Subsequence, LCS) between two texts. It considers the order of words but does not require them to be contiguous.

Thus, ROUGE-N focuses on exact matches of n-grams, while ROUGE-L on more flexible sequence matches, helping to assess the quality of structure and meaning preservation.