Research Of Web Data Mining For Product-innovation

With the rapid development and application of information technology, knowledge economy has begun to lead the development of the world, the competition between enterprises in the new era has changed from competition for resources to product innovation competition. Product is a fundamental survival of enterprises, market is constantly changing, only continuous innovating, developing and manufacturing in line with customer values and aesthetics of products, enterprise development is able to continue. With the rapid development of Web technology such as web portals, search engines and social networks, data appears explosive expansion, how to effectively mining the potential and valuable information from a broad array of Web resources to serve in product innovation has become one of the urgent need to address the current problems.In this thesis, product innovator’s needs as the starting point, we will focus on the Web data mining technology into a study. The main works of the thesis are as following:(1)The problems of search engine in product innovation information acquisition has been analyzed, solutions are proposed to these problems. Through careful studying of HowNet, we expand it to product innovation area, then build Product Innovation concept semantic network, through the Semantic Web finding comprehensive product information, as a parameter to call the search engine, retrieve relevant Web pages、documents and save them to local information cache repository.(2)The current commonly used extraction methods have been Compared and analyzed, then Ontology based extraction method has been selected for the thesis, and the strengths of the Ontology based information extraction method has been researched; the product innovation related Ontology has been constructed, the information extraction algorithm has been researched then evaluated by experiment.(3)The current commonly used text representation models have been Compared and analyzed, the vector space model has been selected for the thesis to represent the extracted information; focusing on the text clustering algorithm, the Jaccard similarity criteria and cosine similarity algorithm have been combined to improve the clustering algorithm, and this algorithm has been used in Web Mining.(4) Based on the above Web data mining methods, a set of product innovation oriented web mining system—PIOWMS has been designed, this system will complete the mining of the closest needs of the product innovators from product innovation Web data, then submit them to the users to server to product innovation.