COMP37332 - Data Integration and Analysis `
Lab test 2: Data mining

Date, time and place

• Tuesday 11 May 2010, 12-14 in the 3rd year lab (Kilburn).

Aims

• Show understanding of the main data mining concepts
• Show familiarity with related software (Weka)
• Analyse the results obtained by data mining

Questions and Marking

• This test is worth 10 marks (overall) and some parts will be assessed during the lab session.
• There will be 5 questions:
  • 2 'theoretical' and
  • 3 practical (involving associations, classification and clustering in Weka).
• You will have 90 minutes to complete 5 questions.
• This is an individual, closed book assignment.
• Slides from the lecture

Datasets

• Dataset1 - vote.arff (used in questions 3 and 4)
• Dataset2 - autos.arff (used in question 5)

Materials and preparation

• Revise the lecture notes for data mininng, association rules, data classification and clustering, and solve all tutorial questions in advance.
  • Data mining introduction slides
  • Association rule mining slides
  • Data classification slides
  • Data clustering slides
  • Tutorial sheet and guide answers
• Familiarise yourself with Weka
  • Weka 3: data mining software
  • Weka tutorial
• Complete all lab tutorials
  • Weka introductory lab/tutorial
    • data files: labor.arff; contact-lenses.arff
  • Association rule mining in Weka (lab/tutorial)
    • data files: contact-lenses.arff; zoo.arff
  • Classification lab/tutorial (Weka)
    • data files: labor.arff; glass.arff; glass-minusatt.arff; glass-withnoise.arff; vehicle.arff
  • Clustering lab/tutorial (Weka)
    • data files: weather.arff; iris.arff; bank.arff flagdata.arff
  • Data Mining tutorial sheet
    • Tutorial Solution