Design and realization of US patent data preprocessing system

In the light of contents, tasks, methods and characteristics of data preprocessing, the universal model of data preprocessing system was presented. Based on that, the data preprocessing system of US issued patent was roundly designed and implemented by improving as well as synthetically applying Chebyshev, optimization processing, association rules etc. diversified algorithms and techniques. The experimental results show that this system triumphantly generates knowledge radix satisfying key data mining algorithms, greatly advancing precision rate, recall rate and efficiency of varied classical mining algorithms.