Overview of computational methods in taphonomy based on the combination of bibliometric analysis and natural language

Abstract

Artificial intelligence tools are new in taphonomy and are growing fast. They are being used mainly to investigate bone surface marks. In order to investigate this subject, a bibliometric study was made to understand the growing rate of this intersectional field, the future, and gaps in the field until now. From Scopus and Google Scholar metadata, graphs were made to describe the data, and inferential statistics were made by regression with the Ordinary Least Squares method. Exploratory analysis with word clouds, topic modeling, and natural language processing with Latent Dirichlet Allocation as a method were also made using the entire corpus from the papers. From the first register until 2023, we found eight articles in Scopus and 32 in Google Scholar; the majority of the studies and the most cited were from Spain. The studies are growing fast from 2016 to 2018, and the regression shows that growth can be maintained in the coming years. Exploratory analysis shows the most frequent words are marks, models, data, and bone. Topic modeling shows that the studies are highly concentrated on similar problems and the tools to solve them, revealing that there is much more to explore with computational tools in taphonomy and paleontology as well.

Key words
artificial intelligence; google scholar; machine learning; paleontology; web scraping; scopus