Study of Corpus Classification Management Method in Construction Correspondence Files

Abstract

The construction correspondence files, including letters, emails and meeting summaries, are used for information exchange during the construction stage. Information about design adjustment, project change and on-site coordination among several parties are included in these files. However, searching for the record is labor intensive and inefficient in the settlement and claim process. Two actual-based corpus classifications were brought out and tested by literature and statistics method in this paper, which could be the guidance for machine language learning. Finally, the application of contract corpus classification is proposed.