PA212 Advanced Search Techniques for Large Scale Data Analytics

Fakulta informatiky
jaro 2023
Rozsah
2/0/0. 2 kr. (plus ukončení). Ukončení: zk.
Vyučující
doc. RNDr. Jan Sedmidubský, Ph.D. (přednášející)
prof. Ing. Pavel Zezula, CSc. (přednášející)
Garance
doc. RNDr. Jan Sedmidubský, Ph.D.
Katedra strojového učení a zpracování dat – Fakulta informatiky
Dodavatelské pracoviště: Katedra strojového učení a zpracování dat – Fakulta informatiky
Rozvrh
Čt 16. 2. až Čt 11. 5. Čt 12:00–13:50 B410
Předpoklady
Knowledge of the basic principles of data processing is assumed.
Omezení zápisu do předmětu
Předmět je nabízen i studentům mimo mateřské obory.
Mateřské obory/plány
předmět má 77 mateřských oborů, zobrazit
Cíle předmětu
The objective of the course is to explain the problems of information retrieval in large collections of unstructured data, such as text documents or multimedia objects. The main emphasis will be given on describing basic principles of distributed algorithms for processing large volumes of data, e.g., Locality-sensitive hashing, MapReduce or PageRank. The algorithms for processing stream data will be introduced as well. The students will also acquire practical experience by applying the presented algorithms to the specific tasks.
Výstupy z učení
After completing the course students are able to:
- Describe algorithmic-based differences between processing offline data collections and online data streams; - Understand the basic principles of distributed algorithms for processing large volumes of data;
- Evaluate the results of algorithms by several metrics;
- Apply presented algorithms, such as K-Means, Locality-sensitive hashing, MapReduce or PageRank, to the specific tasks.
Osnova
  • Introduction – What is searching, Things useful to know
  • Support for Distributed Processing – Distributed file system, MapReduce, Algorithms using MapReduce, Cost model and performance evaluation
  • Retrieval Operators and Result Evaluations – Common similarity search operators, Retrieval metrics
  • Clustering – K-means algorithms, Clustering in non-Euclidean spaces, Clustering for streams and parallelism
  • Finding Frequent Item Sets – Handling large datasets in main memory, Counting frequent items in a stream
  • Finding Similar Items – Applications of near-neighbor search, Shingling of documents, Similarity-preserving summaries of sets, Locality sensitive hashing
  • Searching in Data Streams – The stream data model, Filtering streams
  • Link Analysis – Page Rank, Topic sensitive, Link spam
  • Search Applications – Advertising on the web, Recommendation systems (collaborative filtering), Mining social-network graphs
Literatura
    doporučená literatura
  • P, Deepak a Prasad M. DESHPANDE. Operators for similarity search : semantics, techniques and usage scenarios. Cham: Springer, 2015, xi, 115. ISBN 9783319212562. info
  • LESKOVEC, Jurij, Anand RAJARAMAN a Jeffrey D. ULLMAN. Mining of massive datasets. 2nd ed. Cambridge: Cambridge University Press, 2014, xi, 467. ISBN 9781107077232. info
  • BAEZA-YATES, R. a Berthier de Araújo Neto RIBEIRO. Modern information retrieval : the concepts and technology behind search. 2nd ed. Harlow: Pearson, 2011, xxx, 913. ISBN 9780321416919. info
Výukové metody
Lectures with slides in English. The approach combines theory, algorithms and practical examples.
Metody hodnocení
The final exam consists of only a written part. The student is asked several theoretical and practical questions to verify their knowledge obtained during the course lectures.
Vyučovací jazyk
Angličtina
Další komentáře
Studijní materiály
Předmět je vyučován každoročně.
Předmět je zařazen také v obdobích jaro 2017, jaro 2018, jaro 2019, jaro 2020, jaro 2021, jaro 2022, jaro 2024, jaro 2025.