Raport Badawczy = Research Report ; RB/41/2015

21 pages ; 21 cm

Bibliografia s. 18-21

21 stron ; 21 cm

Bibliography p. 18-21

The paper presents the incorporation of linguistic summaries into the construction of forecasting models. Instead of defining probability distributions, the user validates expected trends and the system creates the probability distributions automatically. Linguistic summaries are constructed as linguistically quantified sentences mined from databases, that may be exemplified by 'Among all increasing trends, most are short'. The proposed approach is evaluated with experiments on real-life time series from the pharmaceutical market and the M3 Competition benchmark datasets.