Abstract:
ERP clients and vendors usually forecast their project interventions very early, before the full specifications for an ERP solution are understood and sometimes before a deal between a vendor / consulting company and a client is concluded. ERP project estimation at the stage of early contractual requirements was difficult as it needs the participation and availability of several experts, each being a specialist in a specific field, e.g. solution architect, technical architect, consultant project manager. Compared with conventional software development approaches, ERP implementations have various kinds of problems. One of these is the development of business transaction documents.
This paper provides a detailed set of effort and schedule prediction models used in the implementation of Stream Serve in Enterprise Resource Planning (ERP) to forecast effort in the production of business transaction documents. The first set of models used to extract expected output of the Business transaction document, which is given by vendor in the Specification (PDF format) and find the complexity of the development. The second set of models used to predict the effort estimation based on complexity of the report. These models are trained by dataset of MOM (M3 Output Management) estimation reports. Finally, System is generating effort estimation to develop Business Transaction Document within few seconds.
