Ghostwriter gatherer: an open source tool for retrieving and scoring web documents from simple search terms

Abstract

The GhostWriter project is an investigation and proof of concept study of automated essay and report writing. Its first component is the search and retrieve tool called the Gatherer,which searches for, retrieves, and scores documents accessible on the Internet via Web tools. Afurther motivation is to test if essays built from gathered material could obtain reasonable scoresfrom automated grading tools such as the MarkIT software with which one of us (RW) isassociated.Web search has immense popular usage, but its efficacy is diminished by duplication and volumeof links, low-quality target material, and commercially driven false leads. We present a relativelysimple but extensible tool that aims to gather web reports or "stories" based on very minimal userinput. We feel that such a tool, even outside of the GhostWriter framework, can aid those whoare preparing position papers in government or corporations, academics seeking to review a fieldthat is not their major focus of research, and students trying to quickly get an overview of asubject. Thus it is of assistance to self-directed learning.