Data mining has drawn much attention in generating the useful information from Web data. Data mining techniques have typically considered quantitative information rather than qualitative, though the qualitative information can often be used to improve the quality of a result. This chapter provides a hybrid data mining application, KBNMiner (Knowledge-Based News Miner), to predict interest rates on the basis of qualitative information on the Web as well as quantitative information stored in a database. The KBNMiner is developed through the integration of cognitive maps and neural networks. To validate the effectiveness of the KBNMiner, an experiment with Web news information is conducted and its results are discussed.