htw saar
Zurück zur Hauptseite

Version des Moduls auswählen:

Machine Learning

Modulbezeichnung: Machine Learning
Studiengang: Praktische Informatik, Bachelor, ASPO 01.10.2011
Code: PIBWI19
SWS/Lehrform: 2V+2U (4 Semesterwochenstunden)
ECTS-Punkte: 5
Studiensemester: 6
Pflichtfach: nein
Arbeitssprache:
Englisch
Prüfungsart:
Klausur
Zuordnung zum Curriculum:
KI575 Kommunikationsinformatik, Bachelor, ASPO 01.10.2011, 6. Semester, Wahlpflichtfach, technisch
PIBWI19 Praktische Informatik, Bachelor, ASPO 01.10.2011, 6. Semester, Wahlpflichtfach, informatikspezifisch
Arbeitsaufwand:
Die Präsenzzeit dieses Moduls umfasst bei 15 Semesterwochen 60 Stunden. Der Gesamtumfang des Moduls beträgt bei 5 Creditpoints 150 Stunden. Daher stehen für die Vor- und Nachbereitung der Veranstaltung zusammen mit der Prüfungsvorbereitung 90 Stunden zur Verfügung.
Empfohlene Voraussetzungen (Module):
PIB115 Informatikgrundlagen
PIB120 Programmierung 1
PIB125 Mathematik 1
PIB215 Mathematik 2
PIB315 Mathematik 3
PIB330 Datenbanken


[letzte Änderung 02.03.2017]
Als Vorkenntnis empfohlen für Module:
Modulverantwortung:
Prof. Dr. Klaus Berberich
Dozent: Prof. Dr. Klaus Berberich

[letzte Änderung 10.02.2017]
Lernziele:
Students know about fundamental supervised and unsupervised methods from machine learning. This includes methods for regression, classification, and clustering. Students understand how these methods work and know how to use existing implementations (e.g., in libraries such as scikit-learn). Given a practical problem setting, students can choose a suitable method, apply it to the dataset at hand, and assess the quality of the determined model. Students are aware of typical data-quality issues and know how to resolve them.

[letzte Änderung 02.03.2017]
Inhalt:
Machine learning plays an increasingly important role with applications ranging from recognizing handwritten digits, via filtering out unwanted span e-mails, to ranking of results in modern search engines. This course covers fundamental supervised and unsupervised methods from machine learning. We will look into how these methods are defined formally, including the mathematics behind them. Moreover, we will apply all methods on concrete datasets to solve practical problems. For this, we will rely on existing libraries (e.g., scikit-learn) that provide efficient implementations of the methods. The course is accompanied by theoretical exercises and project assignments. The former help students to deepen their understanding of the methods; the latter encourage students to solve practical problems by applying what they learnt in the course on real-world datasets.
 
1. Introduction
- What is Machine Learning?
- Applications
- Libraries
- Literature
 
2. Working with Data
- Typical data formats (e.g., CSV, spreadsheets, databases)
- Data quality issues (e.g., outliers, duplicates)
- Scales of measures (i.e., nominal, ordinal, numerical)
- Data pre-processing (in Python and using UNIX commandline tools)
 
3. Regression
- Ordinary least squares
- Multiple linear regression
- Non-linear regression
- Evaluation
 
4. Classification
- Logistic regression
- k-Nearest Neighbors
- Naive Bayes
- Decision Trees
- Neural Networks
- Evaluation
 
5. Clustering
- k-Means and k-Medoids
- Hierarchical agglomerative/divisive clustering
- Evaluation
 
6. Outlook
- Ongoing research
- Competitions (e.g., Kaggle and KDD Cup)
- Other resources (e.g., KDnuggets)


[letzte Änderung 02.03.2017]
Literatur:
P. Harrington: Machine Learning in Action, Manning, 2012
 
G. James, D. Witten, T. Hastie, R. Tibshirani: An Introduction to Statistical Learning - with Applications in R, Springer, 2015
 
A. C. Müller and S. Guido: Introduction to Machine Learning with Python, O´Reilly, 2017
 
M. J. Zaki und W. Meira Jr.: Data Mining and Analysis: Fundamental Concepts and Algorithms, Cambridge University Press, 2014

[letzte Änderung 02.03.2017]
Modul angeboten in Semester:
SS 2017
[Sat Nov 25 08:42:48 CET 2017, CKEY=kml, BKEY=pi, CID=PIBWI19, LANGUAGE=de, DATE=25.11.2017]