Data Mining Strategy of Information Fingerprint Self-learning Mechanism
Download as PDF
Liu Xin, Wang Wenting, Chen Jianfei, Wang Rui, Zhang Hao
Data mining needs various analysis of data, and data preprocessing is needed before all analysis. However, after data cleaning, data integration and data transformation, the data set will still be very large! It will take a long time to analyze and mine complex data directly on massive data, which makes this analysis unrealistic or impractical. Data reduction technology can be used to get the reduction representation of data sets, which is much smaller, but still close to maintaining the integrity of the original data. In this way, mining on reduced data sets will be more effective and produce the same or almost the same analysis results. Through this method, the most representative features are extracted from a large number of features and useful information is analyzed according to needs. With the development of society, traditional security systems based on tokens or passwords are becoming more and more fragile, which can not meet the needs of modern security systems. The fingerprint recognition based on feature extraction comes into being. Two distinct features, endpoint and bifurcation point, are extracted from numerous fingerprint attributes for data mining and analysis.
Data mining, data preprocessing, data reduction, dimension reduction, feature extraction, fingerprint recognition