# One-to-X analogical reasoning on word embeddings: a case for diachronic armed conflict prediction from news texts

## TL;DR

This paper introduces a one-to-X analogy task for word embeddings, applied to predict armed conflict relations from news texts, demonstrating improved accuracy with a thresholding technique and providing a new evaluation dataset.

## Contribution

It extends the word analogy task to a one-to-X format, applies it to conflict prediction, and offers a simple method to enhance diachronic embedding performance.

## Key findings

- Threshold-based filtering reduces false positives.
- Method improves relation prediction accuracy.
- Provides a new dataset for conflict-related analogy evaluation.

## Abstract

We extend the well-known word analogy task to a one-to-X formulation, including one-to-none cases, when no correct answer exists. The task is cast as a relation discovery problem and applied to historical armed conflicts datasets, attempting to predict new relations of type `location:armed-group' based on data about past events. As the source of semantic information, we use diachronic word embedding models trained on English news texts. A simple technique to improve diachronic performance in such task is demonstrated, using a threshold based on a function of cosine distance to decrease the number of false positives; this approach is shown to be beneficial on two different corpora. Finally, we publish a ready-to-use test set for one-to-X analogy evaluation on historical armed conflicts data.