A methodology to allow avalanche forecasting on an information retrieval system

Abstract

This paper presents adaptations and tests undertaken to allow an information retrieval (IR) system to forecast the likelihood of avalanches on a particular day. The forecasting process uses historical data of the weather and avalanche conditions for a large number of days. A method for adapting these data into a form usable by a text‐based IR system is first described, followed by tests showing the resulting system’s accuracy to be equal to existing ‘custom built’ forecasting systems. From this, it is concluded that the adaptation methodology is effective at allowing such data to be used in a text‐based IR system. A number of advantages in using an IR system for avalanche forecasting are also presented.

Citation

Purves, R.S. and Sanderson, M. (1998), "A methodology to allow avalanche forecasting on an information retrieval system", Journal of Documentation, Vol. 54 No. 2, pp. 198-209. https://doi.org/10.1108/EUM0000000007166