Movie Information Web Search and Classification
Frankie Wu · 2008
While there exist several web sites with movie information (cast and crew, etc.), none of these may be use in commercial applications. The same information is dispersed across the world wide web, but in much less predictable formats. This paper presents an investigation of using an existing web search engine to collect HTML pages based on movie titles, and extracting a small subset of the movie information. Classification is performed by machine learning though a maximum entropy Markov model (MEMM) and sequence labeling through a Viterbi decoder.