Persian Named Entity Recognition by Gray Wolf Optimization Algorithm

Named entity recognition (NER) is a subfield of natural language processing (NLP). It able to identify proper nouns, such as person names, locations, and organizations, has been widely used in various tasks. NER can be practical extracting information from social media data. However, the unstructured noisy nature (such grammatical errors typos) causes new challenges for NER, especially low-resource languages Persian, existing methods mainly focus on formal texts English media. To overcome this challenge, we consider Persian an optimization problem use binary Gray Wolf Optimization (GWO) algorithm segment posts into small possible phrases named entities. Later, entities are recognized based their score. Also, prove that even human opinion differ task compare our method with other systems S mathvariant="normal">e mathvariant="normal">p _ mathvariant="normal">T mathvariant="normal">D mathvariant="normal">l 01 dataset results show proposed system obtains higher F1 score comparison methods.