Community intelligence (CI) has developed hundreds of computing algorithms for intelligent data analysis, but still many difficult problems in computer vision, signal processing, or text and multimedia understanding, a problem that requires in-depth learning techniques, is open. Modern data mining package contains many modules for data acquisition, pre-processing, feature selection and construction, sample selection, methods of classification, association and approach, optimization techniques, pattern discovery, clusterization, visualization and post-processing. Operating in the space of all the available data transformation algorithms and optimization techniques using meta-knowledge about the learning process is automatically extracted from a variety of problem-solving experience. Meta-learning shifts the focus from the field overall CI of individual learning algorithms to a higher level of learning how to learn.