TY - GEN
T1 - Does Supervised Learning of Sentence Candidates Produce the Best Extractive Summaries?
AU - Gutiérrez Hinojosa, Sandra J.
AU - Calvo, Hiram
AU - Moreno-Armendáriz, Marco A.
AU - Duchanoy, Carlos
N1 - Publisher Copyright:
© 2020, Springer Nature Switzerland AG.
PY - 2020
Y1 - 2020
N2 - In this work multi-document, extractive summaries have been obtained using supervised learning algorithms in a well-known dataset (DUC 2002); the methodology has three steps: the pre-processing step, which filters irrelevant words and reduces vocabulary using stemming; the representation step, which transforms sentences into vectors; and the classification step which selects sentences for the summary. Noting that the last step is crucial because it determines the relevance of each sentence according to the information included in the embeddings. We found that the classifiers performance is not related to the summary quality mainly classifier’s goal is not aligned to summarizer’s goal, as classifier is based on selecting whole sentences, while summarization is evaluated by n-grams, for example ROUGE-n, and therefore it is relevant while comparing performances between different works in the state of the art.
AB - In this work multi-document, extractive summaries have been obtained using supervised learning algorithms in a well-known dataset (DUC 2002); the methodology has three steps: the pre-processing step, which filters irrelevant words and reduces vocabulary using stemming; the representation step, which transforms sentences into vectors; and the classification step which selects sentences for the summary. Noting that the last step is crucial because it determines the relevance of each sentence according to the information included in the embeddings. We found that the classifiers performance is not related to the summary quality mainly classifier’s goal is not aligned to summarizer’s goal, as classifier is based on selecting whole sentences, while summarization is evaluated by n-grams, for example ROUGE-n, and therefore it is relevant while comparing performances between different works in the state of the art.
UR - http://www.scopus.com/inward/record.url?scp=85092936507&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-60887-3_26
DO - 10.1007/978-3-030-60887-3_26
M3 - Contribución a la conferencia
AN - SCOPUS:85092936507
SN - 9783030608866
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 293
EP - 296
BT - Advances in Computational Intelligence - 19th Mexican International Conference on Artificial Intelligence, MICAI 2020, Proceedings
A2 - Martínez-Villaseñor, Lourdes
A2 - Ponce, Hiram
A2 - Herrera-Alcántara, Oscar
A2 - Castro-Espinoza, Félix A.
PB - Springer Science and Business Media Deutschland GmbH
T2 - 19th Mexican International Conference on Artificial Intelligence, MICAI 2020
Y2 - 12 October 2020 through 17 October 2020
ER -