Analysis of named-entity effect on text classification of traffic accident data using machine learning

<span lang="EN-US">With the rising number of accidents in Indonesia, it is still necessary to evaluate and analyze accident data. The categorization traffic data has been developed using word embedding, however additional work needed achieve better results. Several informative named entities are frequently sufficient differentiate whether or not information on a exists. Named-entities informational characteristics that can offer details about text. influence named-entities thematic text examined this paper. was collected Twitter social media crawl. Preprocessing done at beginning process modify delete useful as well label specified entities. On Support Vector Machine (SVM), scheme comparisons were performed for (i) Word Embedding, (ii) occurrences Named Entities, (iii) combination two known Hybrid. Hybrid produced an improvement classification accuracy 90.27 percent when compared Embedding scheme, according tests conducted 1.885 consisting 788 1.067 non-accident data.</span>