List and describe the five primitives for specifying a datamining task?

Answer 1

1. Task-relevant data: What is the data set that I want to mine? For example: We may specify the number of Computer Science Master and PhD students graduating in fall 2005 and the total number of students graduating in fall 2005. Task-relevant data can be specified by following information: # Name of database to be used: Registrar Office Database, Joe'SS Database. # Name of tables containing the relevant data: Student_Information, Course_Information, CS_Department_Database etc. # Conditions for selecting the relevant data: retrieve data pertaining to graduating students in fall 2005. # The relevant attributes: Student Name, Student Id, Courses Completed, Degree etc. 2. The kind of knowledge to be mined? Association (X = UMR Computer Science students graduating in fall 2005) * major(X, "CS") ^ degree(X, "master") => graduates(X, "fall2005"). * major(X, "CS") ^ degree(X, "phd") => graduates(X, "fall2005"). * major(X, "CS") => graduates(X, "fall2005"). 3. Background knowledge: Concept Hierarchies. The concept hierarchy can be used to define a sequence of mappings from a set of low-level concepts to higher-level, more general concepts. For example, we can define a level schema hierarchy of students who are graduating from UMR as: level < degree < major < school - Phd. < grads < computer science < art and science - Master < grads < computer science < art and science 4. Interestingness Measures. We use the rule: major(X, "CS") ^ status(X, "master") => graduates(X, "fall2005"). Suppose that: - The total number of students in UMR who are graduating in fall 2005 = 100. - The UMR students who are in Computer Science department and graduating with Master's degree in Fall 2005= 10. - The total number of UMR students who are in Computer Science department and graduating in Fall 2005= 15. So, Support = 10/100 = 10% Confidence = 10/15 = 66.6%
5. Presentation and visualization of Discovered Patterns. * major(X, "CS") ^ status(X, "master") => graduates(X, "fall2005"). * major(X, "CS") ^ status(X, "phd") => graduates(X, "fall2005"). * major(X, "CS") => graduates(X, "fall2005"). Presenting using Decision Tree: