Face Extraction from Image based on K-Means Clustering Algorithms
- Yousef Farhang
Usually, data mining is considered as the nontrivial extraction of implicit, previously unknown, and potentially useful information from data. In our data-driven data mining model, knowledge is originally existed in data, but just not understandable for human. Data mining is taken as a process of transforming knowledge from data format into some other human understandable format like rule, formula, theorem, etc. In order to keep the knowledge unchanged in a data mining process, the knowledge properties should be kept unchanged during a knowledge transformation process. Many real world data mining tasks are highly constraint-based and domain-oriented. Thus, domain prior knowledge should also be a knowledge source for data mining. The control of a user to a data mining process could also be taken as a kind of dynamic input of the data mining process. Thus, a data mining process is not only mining knowledge from data, but also from human. This is the key idea of Domain- oriented Data-driven Data Mining (3DM). In the view of granular computing (GrC), a data mining process can be considered as the transformation of knowledge in different granularities. Original data is a representation of knowledge in the finest granularity. It is not understandable for human. However, human is sensitive to knowledge in coarser granularities. So, a data mining process could be considered to be a transformation of knowledge from a finer granularity space to a coarser granularity space. The understanding for data mining of3DM and GrC is consistent to each other. Rough set and fuzzy set are two important computing paradigms of GrC. They are both generalizations of classical set theory for modeling vagueness and uncertainty. Although both of them can be used to address vagueness, they are not rivals. In some real problems, they are even complementary to each other. In this plenary talk, the new understanding for data mining, domain-oriented data-driven data mining (3DM), will be introduced. The relationship of 3DM and GrC, and granular computing based data mining in the views of rough set and fuzzy set will be discussed.