<![CDATA[Text summarization using extraction methods is a technique that summarizes by retaining a subset of sentences to create a summary. There are two types of documents commonly used for summarization: single document and multi-document. Multi-document refers to documents originating from one or more sources that contain several main ideas. The data used in this research is obtained from the E-lapor DIY website, consisting of 1000 data entries. E-Lapor DIY is a website provided by the DIY government to accommodate all public aspirations and complaints, such as damaged roads, broken traffic lights, insufficient street lighting, litter in public places, and more. The accumulation of data and the delayed response time has become an issue for the government in addressing these complaints. This research aims to consider the impact of using feature extraction for text summarization using genetic algorithms. The feature extraction compared in this research is the influence of sentence position in feature extraction. The results obtained show that Precision testing using F1 is 0.64, and without using F1, it is 0.66. Recall testing using F1 is 0.65, and without using F1, it is 0.68. F-Measure testing using F1 is 0.65, and without using F1, it is 0.68. This testing using the algorithm can be an interesting alternative for more time-efficient text summarization.]]>
Copyrights © 2024
