[Data Mining] Klausurthemen

Vorlesungen, Seminare und Praktika aus dem Bereich Daten- und Informationsmanagement
Lectures, seminars and labs from the area Data and Information Management

[Data Mining] Klausurthemen

Beitragvon jsb » 22.02.14 14:57

Hallo,

hat jemand in den letzten Jahren eine Klausur in Data Mining geschrieben?
Kann sich wer erinnern, welche Themen, bzw. Aufgabenstellungen da so dran kamen?

Grüße
jsb
 
Beiträge: 6
Registriert: 04.08.10 14:24
Studiengang: Informatik (B.Sc.)
Studiert seit: SS 10

Re: [Data Mining] Klausurthemen

Beitragvon jsb » 20.03.14 00:37

Folgende Themen waren in der ersten Klausur im WS2013/2014 dran dran (Gedächtnisprotokoll, keine Gewähr für Korrektheit oder Vollständigkeit):

KDD-Process
  • What is the objective of Clustering, Classification, Regression, Frequent Itemset Mining?

Decision Tree Learning
  • Complete a given decision tree using Gini Index

Clustering
  • Hierarchical Agglomerative Clustering:
    Give definitions of Single / Complete / Average Link
    Perform HAC with Single Link and Manhattan Distance on a given point set and draw the dendrogram
  • k-Median: Given a figure of a clustering (k = 2, Euclidean Distance), explain whether this is a valid k-Median Clustering or not
  • DBSCAN:
    Given a figure of a clustering (and DBSCAN parameters), explain whether this is a valid DBSCAN Clustering or not
    Given a figure of a point cloud, determine DBSCAN parameters minPts, epsilon using the heuristic

FP-Trees
  • Given an initial FP-Tree, complete the FP-Tree Growth algorithm. Determine all Frequent Itemsets and draw all conditional FP-Trees.

B-Trees
  • Given the coordinates of some 2D points, determine their index on a 2nd order Hilbert Curve
  • Given a B-Tree (M = 2), insert some keys
  • Given a B-Tree (M = 4), delete some keys

Bitmap Index
  • Given a set of values, determine their representation in a Bitmap Index using: Simple Encoding, Range Encoding, Interval Encoding
  • Formulate some queries for these three Bitmap Indices

Bayes Classifier
  • Give the definition of the Bayes Classifier
  • Give the definition of the Naive Bayes Classifier
  • Classify an object using a Naive Bayes Classifier
  • Give an example where the Bayes Classifier performs better than the Naive Bayes Classifier

Prediction
  • Given a training set and a test set of 2D points.
  • Train a constant model using the training set and determine the Mean Average Deviation using the test set
  • Train a Regression Tree with one split (by just looking at the points and drawing) using the training set and determine the Mean Average Deviation using the test set

Bias / Variance
  • Give definitions of Bias and Variance
  • Explain the relation between Bias, Variance, Overfitting

Distance Functions
  • Given a figure with three points, answer some questions about their L1 and L2 distances to each other
  • Prove or disprove whether some given distance functions are metrics
jsb
 
Beiträge: 6
Registriert: 04.08.10 14:24
Studiengang: Informatik (B.Sc.)
Studiert seit: SS 10


Zurück zu Daten- und Informationsmanagement / Data and Information Management