POMon: From Simple Keyword Matching to Stream-Based Live Probabilistic Topic Matching
Kun Ma, Xuewei Niu, Ziqiang Yu, Ke Ji · 2018
Public opinion monitoring is defined as story detection and matching with particular Internet news event. Traditional method used topic discovery and text matching techiques to get results. However, it has some limitations: hidden topic discovery and incorrect match ranking on big data. In this demonstration, we present the architecture evolution of public sentiment monitoring system (POMon) and ranking method of live probabilistic topic matching. Meanwhile, we discuss the evolution from simple keyword matching to stream-based live probabilistic topic matching.