Classification Model Induction Based on User Preferences

摘要

The data mining has been influential in gathering information and data in the work of organizations and business, medical, engineering etc. The strength of data mining is the help in the matter of technique in information management. This technique can be used to find important information along with other information. This paper proposes highlight of the classification for the data in each group to work in the field of medicine that is useful for clinicians and patients that may communicate through the model made from data mining, such as "How much the amount of sugar in the blood of patients to be a risk of diabetes?". It can be output in the form of decision trees. But when we want to know just some of the information, a whole decision tree is superfluous. This research has focused on this point of knowledge reduction. We will use the method of logic programming to imitate the functionality of data mining to extract patterns from data taken from real sources. The results of pattern extraction will be in the form of rules. We increase efficiency of knowledge navigation by allowing users to specify constraints or preferences, which will help in the selection of specific rules of interest. Our methodology can enhance the search for answers, as well as reduce the time to locate all the rules.