Proper Nouns Recognition in Arabic Crime Text Using Machine Learning Approach

Named Entity Recognition (NER) identifies proper nouns in a text and categorizes it as a distinct kind of named entities. This function enables the extraction of peoples name, locations, organizations, and currencies. Several research abound in this area in Arabic NER is concerned. However, recognizing Arabic named entities is challenging due to the complexity in the Arabic language. These complexities are represented by non-existence of capitalization feature which facilitates the process of NER. Furthermore, there is a lack of lexical corpora that may include all the Arabic NEs. On other hand, most of the approaches that have been proposed for Arabic NER were based on handcrafted rule-based methods which can be laborious and time consuming. Therefore, this paper presents our attempt at recognizing and extracting the most important named entities, such as names of persons, locations, organizations, crime types, dates and times in Arabic crime documents using the Decision Tree classifier and feature extraction for crime dataset. The dataset consists of varying data sizes collected from online resources and undergone multiple pre-processing tasks. Additionally, the feature extraction task which includes POS tagging, keyword trigger, definite articles and affixes has been performed. Furthermore the classifier will utilize these features in order to classify the named entities. The results demonstrate that the use of the Decision Tree (DT) yields good results in small size datasets, but failed when the dataset is large in the case of the crime domain. This is due to the difficulty of relevant keywords and features in such fields. The best result for this experiment is 81.35% F-measure.