.png){kind=icon}
.png){kind=icon}
Return
A binary grey wolf optimizer to solve the scientific document summarization problem
.png){kind=icon}
PRE
AI
DOI：10.1007/s11042-023-16358-x.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The extraction of information from the extensive volume of online textual data poses a significant challenge, and text summarization plays a pivotal role in overcoming this challenge. Conventionally, an extractive text summary, which consists of the most relevant sentences from the text itself, can effectively represent the given text. However, identifying such a subset of sentences is challenging. To overcome this problem, this paper introduces a Binary Gray Wolf Optimization (BGWO)-based text summarization approach that tackles the sentence selection problem. The proposed system performs pre-processing on the input texts, identifies noteworthy features, and generates a text segment that includes the most relevant sentences. Subsequently, the BGWO algorithm is employed to generate an optimal summary from the text segment. The BGWO-based text summarization approach begins by initializing the population as a set of feasible solution vectors represented by binary values, indicating the presence or absence of sentences in the summary. Thereafter, fitness functions incorporating textual features are constructed, and the population's fitness is evaluated. Through non-dominated sorting and crowding distance, individuals are categorized as alpha, beta, delta, or gamma wolves. In each iteration, their positions are updated using crossover and mutation operations, and individuals are ranked based on their fitness scores. Finally, the alpha wolf is selected as the optimal summary candidate. The proposed method is evaluated and compared to various existing methods on the DUC-2001, DUC-2002, and ScisummNet datasets using ROUGE measures. Statistics based on ROUGE scores are also computed, demonstrating that the proposed system is statistically significant. The experimental results demonstrate that BGWO outperforms ROUGE scores for single-document summarization, including low-resource documents.
Keywords:
ScisummNet
Grey wolf optimization
Extractive text summarization
News summarization
Scientific document summarization
Research article summarization
Journal
IF:
3
Papers:
1.9W
Citations:
3.2W