Development of Automatic Keirin Article Generation System

DOI

Bibliographic Information

Other Title
  • 競輪における予想記事の自動生成システムの開発

Abstract

The burden of manually creating Web content is high, and there is a high demand for automatic generation technology. In this paper, we developed a system to automatically generate articles for Keirin. Our system predicts race results including race development using Transformer, and generates articles by a rule-based method. Experiments on dataset show that our predictor performs better than the odds and that it is possible to make a profit by selecting tickets based on the predictor's output. The system-generated articles achieved the same or better score as the existing articles in human evaluation.

Journal

Keywords

Details 詳細情報について

Report a problem

Back to top