Fine-grained Classification of A Million Life Trajectories from Wikipedia

TL;DR

This paper introduces a large-scale method for classifying detailed life activities of notable individuals from Wikipedia, using syntactic graphs and large language models to improve accuracy and create a comprehensive dataset for human dynamics research.

Contribution

The study presents a novel approach combining syntactic graphs and LLMs to classify fine-grained life activities from Wikipedia, resulting in the largest dataset of its kind.

Findings

Achieved 84.5% classification accuracy

Constructed a dataset with 3.8 million labeled activities

Surpassed baseline methods in activity classification

Abstract

Life trajectories of notable people convey essential messages for human dynamics research. These trajectories consist of (\textit{person, time, location, activity type}) tuples recording when and where a person was born, went to school, started a job, or fought in a war. However, current studies only cover limited activity types such as births and deaths, lacking large-scale fine-grained trajectories. Using a tool that extracts (\textit{person, time, location}) triples from Wikipedia, we formulate the problem of classifying these triples into 24 carefully-defined types using textual context as complementary information. The challenge is that triple entities are often scattered in noisy contexts. We use syntactic graphs to bring triple entities and relevant information closer, fusing them with text embeddings to classify life trajectory activities. Since Wikipedia text quality varies, we use LLMs to refine the text for more standardized syntactic graphs. Our framework achieves 84.5\% accuracy, surpassing baselines. We construct the largest fine-grained life trajectory dataset with 3.8 million labeled activities for 589,193 individuals spanning 3 centuries. In the end, we showcase how these trajectories can support grand narratives of human dynamics across time and space. Code/data are publicly available.