With the advent of the Internet age, the network generates a lot of product evaluation information. This information not only provides valuable reference information for potential buyers, but also gives these product manufactory the direct feedback from product users. These help producers to better understand their product and improve it.The method of product feature extraction and analysis can be divided into supervised machine learning methods, semi-supervised machine learning algorithms and unsupervised machine learning algorithm. For the lower requirement of labeled information and pretty good result, semi-supervised machine learning algorithms become the mainstream method of Product feature extraction and analysis. In this paper, a novel approach for lexical acquisition based on Bootstrapping has been presented. And make a little attempt on unsupervised machine learning algorithm.In this paper, we set a new product feature extraction and analysis system whose core is based on MSGA-Bootstrapping. This system only needs a few seed words and amount of unlabeled corpus to extract product feature.Experiments show that the proposed algorithm in this paper can be a effective solution to product feature extraction and analysis. |