SMU Data Science Review
Abstract
In this paper, an analysis is presented of the changing baby namespace and a model is created for predicting if a name's popularity is trending up or down. Just as cultures and societies change over time, baby names evolve to reflect these changes. By analyzing name phonemes and historical influences, one can better understand the underlying causes of the changing name trend. Utilizing the U.S. Social Security Administration (SSA) name registry and historical figure data sets, the influence of historical figures and name pronunciation on the naming trend was examined. Two neural networks were created to predict name trend, one utilizing name count and the other utilizing name pronunciation. Phoneme embeddings were also created to cluster and visualize similar and dissimilar sounding names. The analysis concluded that while historical factors do influence the U.S. naming trend, these factors are too inconsistent and sporadic to include in a name forecasting model. The phoneme-driven model classified name trend with 72% percent accuracy, while the model using name counts achieved 92% accuracy. Based on these results, there is a relationship between similar sounding names and their popularity trends, but it is not as predictive as purely using name count.
Recommended Citation
Ludwig, Laura; Hightower, Mallory; Engels, Daniel W.; and McGee, Monnie
(2019)
"Analyzing Influences on U.S. Baby Name Trends,"
SMU Data Science Review: Vol. 2:
No.
3, Article 8.
Available at:
https://scholar.smu.edu/datasciencereview/vol2/iss3/8
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial 4.0 License
