B.Sc. (Hons.) Computer Science(w.e.f.

2011)

CSHT 616 (IV) Data Mining Guidelines

Sr. No. of
Units Topics Chapter Reference
No. Lectures
1.1 - Why Data Mining? 1.2 - What Is Data Mining?,
1. Introduction 1 1 2
1.4 - What Kinds of Patterns Can Be Mined?
2.1- Data Objects and Attribute Types, 2.2 - basic
2. Data 2 1 2
Statistical Descriptions of Data (upto page 51)
Preprocessing
3. 2.2 – Data Quality, 2.3 – Data Preprocessing 2 2 6
Association 6.1 – Basic Concepts,6.2- Frequent Itemset Mining
4. 6 1 12
Analysis Methods (upto page 259)
4.1 – Preliminaries, 4.2 – General Approach to Solving
a Classification Problem, 4.3 Decision Tree Induction
5. 4 2 7
(Till Pg. 165), 4.5 – Evaluating the Performance of a
Classifier
Classification
5.1 – Rule Based Classifier (upto page 212),5.2 –
Analysis
Nearest Neighbor Classifiers, 5.3– Bayesian Classifiers
6. (Complete for discrete data and only introduction of 5 2 8
Bayes classifier for continuous attributes) till pg. 233,
5.7.1 – Alternative Metrics
10.1- Cluster Analysis, 10.2 - Partition Methods, 10.3 -
Cluster
7. Hierarchical Methods (uptopg 462), 10.4 - Density 10 1 11
Analysis
Based Methods (uptopg 473)

Course Books:

1. Data Mining: Concepts and Techniques, 3nd edition,Jiawei Han and MichelineKamber
2. Introduction to Data Mining, Pang-Ning Tan, Michael Steinbach, Vipin Kumar, Pearson
Education.

References:

3. Data Mining: A Tutorial Based Primer, Richard Roiger, Michael Geatz, Pearson Education 2003.
4. Introduction to Data Mining with Case Studies, G.K. Gupta, PHI 2006
5. Insight into Data mining: Theory and Practice, Soman K. P., DiwakarShyam, Ajay V., PHI 2006
Practical List: Practical are to be done using Weka, and a report prepared as per the format*. The
operations are to be performed on built-in dummy data sets of weka and/or the downloadable datasets
mentioned in references below. Also wherever applicable, the parameter values are to be varied (upto
3 distinct values). The 'Visualize' tab is to be explored with each operation.

1. Preprocessing : Apply the following filters

weka>filter>supervised>attributed>

 AddClassification
 AttributeSelection
 Discretize
 NominalToBinary

weka>filter>supervised>instance

 StratifiedRemoveFolds
 Resample

weka>filter>unsupervised>attribute>

 Add
 AddExpression
 AddNoise
 Center
 Discretize
 MathExpression
 MergeTwoValues
 NominalToBinary
 NominalToString
 Normalize
 NumericToBinary
 NumericToNominal
 NumericTransform
 PrincipalComponent
 RandomSubset
 Remove
 RemoveType
 ReplaceMissingValues
 Standardize

weka>filter>unsupervised>instance>

 Normalize
 Randomize
 Standardize
  RemoveFrequentValues
 RemoveWithValues
 Resample
 SubsetByExpression

2. Explore the 'select attribute' as follows

weka>attributeSelection>

 FilteredSubsetEval
 WrapperSubsetEval

3. Association mining

weka>associations>

 Apriori
 FPGrowth

4. Classification**

weka>classifiers>bayes>

 NaïveBayes

weka>classifiers>lazy>

 IB1
 IBk

weka>classifiers>trees

 SimpleCart
 RandomTree
 ID3

5. Clustering**

weka>clusters>

 SimpleKMeans
 FarthestFirst algorithm
 DBSCAN
 hierarchicalClusterer
 *Prescribed format:

Dataset Task Algorithm Filter Parameters Observations Inference Remarks

For all attributes,

The dataset has
Unsupervised -> mean =0 and
Iris Preprocessing ignoreClass: false been
Standardize standard deviation
standardized
=1

The minimum and

The dataset has
Unsupervised -> scale: 1.0, maximum of all
Iris Preprocessing been normalized
Normalize translation: 0.0 attributes is 0 and
in [0,1]
1 respectively

The minimum and

The dataset has
Unsupervised -> scale: 2.0, maximum of all
Iris Preprocessing been normalized
Normalize translation: 0.0 attributes is 0 and
in [0,2] range
2 respectively

** Proper graphs are to be drawn to compare the accuracies achieved by the variations mentioned below.

 Applying different algorithms to the same dataset.

 10%, 20%, 30%, 40% and 50% Noise.
 Applying different datasets to the same algorithm.

References for the data sets to be used for the experiments:

1. http://archive.ics.uci.edu/ml/
2. http://www.kdnuggets.com/datasets/index.html
3. https://wiki.csc.calpoly.edu/datasets/wiki/apriori (for Association Mining)