Statistical Features and PageRank Scoring Fusion for Arabic Text Summarization

online text content is increasing daily, making it a big challenge for readers to find the required information. So, providing a gist of the available text content can help those readers save their time and effort. This operation is completed by Automatic Text Summarization (ATS), which creates a summary for the entire text. Automatic Text Summarization can be achieved by many techniques, such as graph theory, statistics, machine learning, and clustering methods. Each of the above techniques takes care of a specific issue of text which other techniques do not. Text summarization for the Arabic language suffers from a lack of resources and language complexity. So, it is worthwhile to take the advantages of two or more text summarization techniques and fuse them together to improve their performance and enhance their summary quality. This study proposes a method that combines two of the ATS methods, which are the statistical based method and the graph-based modified PageRank algorithm. Graph-based techniques take care of relationships among document sentences, whereas statistical-based techniques take care of signalizing important sentences. Experiment results show that proposed method achieved a good performance. The Essex Arabic Summaries Corpus (EASC) was used as evaluation data.