Combining clustering and a decision tree classifier in a forecasting task
Arnis Kiršners, Sergei Parshutin, Arkady Borisov · Automatic Control and Computer Sciences · 2010
A joint analysis of continuous (time series demand observations) and discrete (well-describing parameters) data is studied. Such data mining techniques as data collection, preprocessing, clustering analysis, and classification are considered. Upon continuous data preprocessing and clustering, images of possible sales development are constructed. A new product’s demand is searched for using inductive decision trees built on well-describing data.