A Single-Station Weather Forecasting Expert System.

Publisher Website
Google Scholar
Abstract
This report describes the development of an early prototype expert system for the classical single-station weather forecasting problem. The skills required to maximize forecast quality based on such limited data are being lost though they continue to have application to certain military scenarios which include communications interruptions. Knowledge for the expert system was obtained from a meteorologist experienced in single-station weather forecasting by working through case study data. The prototype was developed on a microcomputer and uses forward and backward chaining rules, frames, and other functions and procedures written in LISP. The expert, system is able to emulate the expert for a several day case study used in the development, although much more work must be done to make the system useful for the general meteorological situation.

This publication has 0 references indexed in Scilit: