Academic year2016/17
CourseCOMPUTER ENGINEERING
Code586II
Credits9
PeriodSemester 1
LanguageEnglish
Modules | Area | Type | Hours | Teacher(s) |
BUSINESS INTELLIGENCE | ING-INF/05 | LEZIONI | 90 | |
Programma non disponibile nella lingua selezionata
Knowledge
The students who successfully complete the course will have a solid knowledge of the main techniques used in data preprocessing, data warehouse, frequent pattern mining, classification, prediction, clustering and outlier detection. This knowledge will allow them to tackle each type of data mining problem and to identify the most suitable technique for solving it.
Assessment criteria of knowledge
During the oral exam the student must be able to demonstrate his/her knowledge of the course material and be able to identify the most suitable solutions for specific data mining problems.
Methods:
- Final oral exam
- Written report
Further information:
The student is requested to implement one algorithm of data mining in Weka and experiment this algorithm. Weka is a public domain software written in Java. The written report summarises the results obtained by the experimentation performed by the student.
Teaching methods
Delivery: face to face
Learning activities:
- attending lectures
- individual study
- Laboratory work
- Practical
Attendance: Advised
Teaching methods:
- Lectures
- Seminar
- Task-based learning/problem-based learning/inquiry-based learning
- laboratory
- project work
Syllabus
Data Preprocessing: data cleaning, integration, reduction, transformation and discretization.
Data warehouse: basic concepts, conceptual models, design, implementation and usage, data cube computation methods and online analytical processing, multi-dimensional data analysis in cube space.
Frequent pattern mining: basic concepts, A-priori algorithm, Pattern-Growth approach, vertical data format, pattern evaluation methods, constraint-based frequent pattern mining.
Classification: basic concepts, decision tree induction, Bayes classification methods, rule-based classification, lazy learners, techniques for improving accuracy, model evaluation and selection.
Clustering: basic concepts, partitioning methods, hierarchical methods, density-based methods, grid-based methods, model evaluation and selection.
Outlier detection: statistical, proximity-based, clustering-based and classification-based approaches.
Bibliography
Recommended book:
J. Han and M. Kamber. Data Mining: Concepts and Techniques. Morgan Kaufmann, 3rd ed., 2011
Papers on the different algorithms described during the course
Slides of the lectures
Updated: 14/11/2016 17:27