Evidence record · UDINUS · SDGs
- Unit
- Rektorat UDINUS
- Period
- 2026
- Goals
- Record
- ec251785-1960-4247-b8f6-961e774b877e
- Published
- 2026-10-30
Universitas Dian Nuswantoro conducted an activity to develop a Named Entity Recognition model for identifying human hobbies from Indonesian-language tweet data. The activity involved UDINUS researchers Nurchim, Muljono, Edi Noersasongko, and Ahmad Zainul Fanani. The source presents hobbies as a component of human identity that can provide social and emotional satisfaction. It also explains that people express viewpoints about their pastimes through social media platforms. Twitter is described as a social media service that allows users to disclose personal information, making the information available for analysis through Named Entity Recognition.
The activity used a manually categorized entity-tag model. The development process consisted of three stated steps: inputting data sources, creating the Indonesian Named Entity Recognition model, and evaluating the model. The model was created manually using the Python platform and the Spacy library. Testing produced fifteen categories of label tags and an average F1 score of 92,31%. The source further states that models for other sentence patterns require additional study because each part of the dataset has a different sentence pattern.
Through this activity, UDINUS presents research focused on developing and evaluating an Indonesian-language computational model for identifying human hobbies in tweet data. The use of Python and Spacy, together with manual model creation, entity-tag categorization, and evaluation, frames the activity as research involving computational methods and model development. The stated need for further study identifies sentence-pattern variation as a continuing consideration in this research.