Behind the Name: A Comparative Framework for Age Estimation of Taiwanese Names

Demographic information provides valuable insights for understanding the personalities and behaviors of individuals. With the rise of privacy concerns, several types of demographic information are not publicly available on both online and offline platforms, even basic information, such as age and gender. This study commits to providing an estimation of age and gender information, based on a basic demographical element, which is the name of a person. Regarding a name itself, it implicitly carries various meanings and clues through the naming process, such as parental expectations and significance related to era or gender. Facing the scarcity of name datasets, this research collected a large number of Taiwanese names on both online social media and offline resources. By probing Taiwanese culture on naming, such as the fortune-telling customs in Asia, we designed comprehensive features and represented the meaning of names by word embedding to capture their hidden semantic expectations. A two-step comparative framework was further applied to reduce the prediction error. The experimental results show that our approach can successfully estimate both age and gender from only a name with the designed features from Taiwanese culture.