| TANA48 | Numerical Methods for IT Applications , 6 ECTS credits. /Numeriska metoder för IT-tillämpningar/ |
|||
OBS! |
Overlapping course contents: TANA25. |
|||
Prel. scheduled
hours: 38 |
||||
|
Area of Education: Science Subject area: Mathematics, applied mathematics |
||||
| Advancement level
(G1, G2, A): A
|
||||
| Aim: With IT applications we understand primarily data mining and patter recognition. Student should acquire knowldege on basic concepts in the area and be acquainted with selected algorithms. After the course the student is expected to be able to 1) use the singular value decomposition (SVD) and similar matrix factorizations to solve least squares problems and compute orthogonal bases, 2) use SVD, clustering and similar methods to do text mining, pattern recognition and computation of pagerank, 3) use simple software (parser) for text processing for text mining |
||||
| Prerequisites: (valid for students admitted to programmes within which the course is offered) Basic courses in numerical algorithms (scientific computing) and programming. Note: Admission requirements for non-programme students usually also include admission requirements for the programme and threshhold requirements for progression within the programme, or corresponding. |
||||
| Organisation: The theory is presented mainly at seminars. Programming projects will give practical experience in solving applied problems and using suitable tools. |
||||
| Course contents: Orthogonal transformations, QR decomposition, least squares problems, recursive least squares. Eigenvalue and singular value decomposition (SVD), separation of signal and noise. Data compression by SVD. Factorization of tensors. Algorithmic aspects of matrix factorizations (robustnessm efficiency, program libraries), the use of the algorithms as building blocks for solving applied problems. Pattern recognition (handwritten digits, face recognition). Image restoration. Information retrieval and search engines, text-mining. Clustering and classification. |
||||
| Course literature: L. Elde'n: Matrix methods in data mining and pattern recognition, SIAM 2006 |
||||
| Examination: | ||||
| Written examination. Laboratory work. |
2 p 2 p |
/ / |
3 ECTS 3 ECTS |
|
Course language is English. Department offering the course: MAI. Director of Studies: Tommy Elfving Examiner: Lars Eldén Link to the course homepage at the department Course Syllabus in Swedish |
||||
|
|
 |
|||||
| ||||||