51 Application of an Integrated Approach Based on Zoning Search to Sentiment Analysis

To acquire hot features in detecting sentiment or opinion from the Web, the paper puts forward an integrated approach using self-adaptive genetic algorithm for clustering based on zoning and global search for the largest condensation points on the premise that the entire data space has been divided dynamically and density points in each partition has been calculated. It is proved that the proposed method could not only extract the hot features clearly from nearly 400 documents related to pornography, violence, environment and estate, but also take on good performance by a simulation experiment.