Incremental learning of aspect model on streaming documents
Te-Min Chang, Wen-Feng Hsiao, Cheng-Wei Wu · 2010
This research is to propose an IR related technique, the incremental aspect model (ISM), which not only uncovers latent aspects from the collected documents but also adapts the aspect model on streaming documents chronologically. ISM includes two stages: in Stage I, probabilistic latent semantic indexing (PLSI) technique is used to build a primary aspect model; and in Stage II, with out-of-date data removing and new data folding-in, the aspect model can be expanded using the derived spectral method if new aspects significantly exist. Two experiments on text clustering tasks are conducted accordingly. Results show the ISM has robust performance in terms of its incremental learning ability.