<link rel="stylesheet" href="styles.cf390c6bfbe7555f.css">

AgriQuant: An ontology-backed event corpus for reproducible event studies in animal-health and agribusiness equities

Loading...
Thumbnail Image

Date

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

An event study asks how a company's stock moved around a piece of news, such as a drug approval or an acquisition. The statistics are well established, but the list of events behind them is often hard to check, and a wrong date or a duplicated record can change the result even when the code is correct. This research presents AgriQuant, a knowledge-graph-backed system that keeps financial events for animal-health and agribusiness companies in a form that can be checked, traced, and rerun. The events are encoded as a Resource Description Framework knowledge graph, the Web Ontology Language defines the kinds of events, and the Shapes Constraint Language checks that every record is complete. Each event keeps two dates, when it happened and when the public first learned about it, and only events with a confirmed public date are analyzed. A review of primary sources gave every collected record a final decision and produced a corpus of 162 events across 18 companies. A case study applies a standard market-model event study to the eight animal-health stocks among them from 2021 to 2025, using a fixed snapshot of daily closing prices from Stooq. Rerunning the analysis reproduces the saved results exactly, every result lists the events behind it, and the sensitivity checks show how much the sample and the results depend on which events count. The outcome of this research is not a stock forecast but a reproducible link from each reviewed event and its date evidence to the exact data and result it produced, which makes event-study results traceable and explainable.

Description

Keywords

event study, knowledge graph, ontology, reproducibility, animal health, agribusiness

Graduation Month

December

Degree

Master of Science

Department

Department of Computer Science

Major Professor

Hande McGinty

Date

Type

Thesis

Citation

Endorsement

Review

Supplemented By

Referenced By