XML Transformation-based three-stage pipelined Natural Language Generation System

The purpose of this study is to generate weather forecasts from XML-stored weather dataset. Reiter and Dale (2000) proposes Natural Language Generation (NLG) System as three-stage pipelined modularized system. We follow them, but not only regard our input data as XML database, but implement three module as XML format transformation process. As a result, new insights at the research eld of report generation from database (Kittredge and Polguere, 2000) are illustrated.

• Type: Conference
• Year: 2001
• Where: NLPRS
• Authors: Yohei Seki
• Keywords: Natural Language Generation, Natural Language Processing, NLPRS 2001, XML Format Transformation, XML-stored Weather Dataset