Project in String Processing Algorithms
2
Algorithms and machine learning
Advanced studies
Implementation and experimental comparison of string processing algorithms, and presentation of the results.
| Year | Semester | Date | Period | Language | In charge |
|---|---|---|---|---|---|
| 2017 | spring | 17.01-28.02. | 3-3 | English | Juha Kärkkäinen |
Lectures
| Time | Room | Lecturer | Date |
|---|---|---|---|
| Tue 12-14 | B119 | Juha Kärkkäinen | 17.01.2017-28.02.2017 |
General
NOTE: The opening meeting is on the second week Tue 24.1. at 12-14 in B119.
The project consists of
- implementation of one or more string processing algorithms
- experimental comparison and/or analysis of the algorithm(s)
- presentation of the results
The project can be done in groups of at most three students. In a group each student is responsible for specific algorithms and the group together is responsible for the experiments and the presentation.
The course assumes String Processing Algorithms -course or similar knowledge as background.
Completing the course
Algorithm implementation
The algorithms can be implemented with any programming language under the restriction that the programs can be compiled and executed on the department computers.
Experiments
The purpose of the experiments is to determine how the performance of the algorithms changes with different inputs, different parameters settings, different algorithms etc.. An important part is choosing the test data.
Poster
The results of the experiments are presented as a poster. There will be an open poster presentation session on the last week, where other students and staff of the department can come to view the posters and ask questions.
Grading
Each part of the project (implementation, experiments, presentation) contributes one third to the total score. In general,
the experiment and presentation score will be the same for all members of a group. The implementation scores will be personal.
Literature and material
Lectures
Poster
- Example poster (incl. scripts for printing the poster): example_poster.zip
Test data collections
- "smart" corpus: small files (~5 MiB). Mostly natural text.
- Pizza & Chili corpus: standard and highly-repetitive testfiles. Medium testfiles (50MiB - 1GiB). Various types of data.
- Wikipedia dumps: large files (usually >50 GiB). Natural text + XML.
These are only examples. The experiments could also include other data, e.g., worstcase inputs.