Highly Relevant Routing Recommendation Systems for Handling Few Data Using MDL Principle
share
Many classification algorithms existing today suffer in handling many so-called "few data" or "limited data" instances. In this paper we show how to score query relevance with handle on few data by adopting a Minimum Description Length (MDL) principle. The main outcome is a strongly relevant routing recommendation system model (average(F)=0.72, M>=0.71; all scales of 1) supported by MDL based classification which is very good in handling few data by a large percentage margin of data degeneration (up to 50 percent loss).
READ FULL TEXT