Design and Implementation of a Natural Language Chat System Based on Statistical Learning

Asan experiment in natural language processing,via learning from huge corpus based on the optimized multi-order Markov models,we built a language statistical model.This model can automatically create the answer-text,and can achieve topic detection and track based on the extended and an optimized TF-IDF algorithm.A chat system,with a desktop version and an on-line version,was respectively implemented based on a free text in an Open-Domain.