A new web information extracting method based on multi-coordinate

To sovle the problems of lower accuracy and higher re-build workload caused by single path-based data locating methods in the traditional Web information extracting, a new method called multi-coordinate locating of information extracting has been presented in the paper, which constructs three different coordinate systems such as global coordinate, local coordinate and random coordinate to locate the information in HTML page. And it is able to re-locate the path of data by the self-restoring of wrapper based on multi-coordinate systems when the HTML document changes. Each of the three coordinate locating methods has been described in detail. By developing a prototype system and doing some experiments, it can be proved from the results that the multi-coordinate method can improve the tolerance to Web page´s changes of the wrapper without adding extra costs and decreasing its performance.