Data Mining

Data Mining

1. Mining techniques

   • Classification
     • Binary Classification problem
       • predict if a patient is high-risk or low-risk at emergency room
       • decide if a credit card application should be approved or not.
     • Given a set of instances (tuples, examples, data records, cases), as a training set, each described by attributes and a class. Learn a classification model from the training set that can be used to predict the classes of new/future instances.
     • Types of learning
       • Supervised learning
         • Supervision: the data is labelled with pre-defined classes.
         • 2-step process
           • 1. Learning/Training: learn a model with a learning algorithm using the training data
             • 2. Testing: test the model using unseen test data to assess the model performance. Accuracy = (# of correct classifications / # of all test cases)
       • Unsupervised learning
         • Class labels of the data are unknown. Given a set of data, the task is to establish the existence of classes, clusters in the data
     • Approaches
       • [*] Decision trees
         • [*] Naive Bayesian
           • Bayesian belief network
             • Rule based classifiers
               • Neural networks - backpropagation
                 • Support Vector Machines
                   • [*] Associative classification - classification by association rule analyis
                     • K-nearest neighbor classifiers

2. Data

   • Data - collection of Instances and their Attributes. A collection of attribute values describe an instance
   • Types/Domains of attributes
     • Numerical: domain is ordered and can be represented on the line (e.g. age, income)
     • Nominal/Categorical: domain is a finite set of name values without any specific order (e.g. color, postal codes, race, ID, martial status)
     • Ordinal: domain is ordered but absolute differences between values is unknown (e.g preference scale from 1 to 10, severity of an injury, grades, height - tall, medium, short).

3. Information Theory

   • provides a mathematical basis for measuring the information content.
   • If one already has a good guess about the answer, then the actual answer is less informative