News e Bacheca Elettronica

Breve Corso su Machine Intelligence and Granular Mining: Relevance to Big Data

Valutazione attuale: 0 / 5

Stella inattivaStella inattivaStella inattivaStella inattivaStella inattiva


Machine Intelligence and Granular Mining: Relevance to Big Data
Sankar K. Pal,
Center for Soft Computing Research, Indian Statistical Institute, Kolkata
Tuesday Nov 21, 2017 3PM-5PM, AULA 4
Thursday Nov 23, 2017 3PM-5PM, AULA 4

La frequenza del Corso darà dirittto al riconoscimento di 1 CFU.

Syllabus -
A common way to speed up an application is to break it up into smaller tasks that can run concurrently. Granular computing will permit much smaller tasks than before, and many more of them; this will allow Big Data applications to incorporate large amounts of data into sophisticated machine learning algorithms, while still completing quickly enough to respond to devices in real time.
The course has two parts. First it describes the:
• Role of pattern recognition in data mining and machine intelligence
• Features of granular computing
• Significance of fuzzy sets and rough sets in granular computing
• Characteristics of fuzzy sets and rough sets in handling uncertainties arising from overlapping regions/ concepts, and granularity in domain respectively   Relevance of defining the generalized rough sets and entropy by embedding fuzziness into rough sets; providing a stronger paradigm for uncertainty modeling
The second part deals with various mining applications such as in:
• Video tracking in ambiguous situations
• Bioinformatics (e.g., selection of miRNAs for cancer detection)
• Social network analysis (e.g., community detection)
The applications demonstrate the roles of different kinds of granules, rough lower approximation, and various information measures. Granules considered range from crisp, fuzzy, 1-d, 2-d and 3-d to regular shape and arbitrary shape. While the concept of rough lower approximation in temporal domain provides an initial estimate of object model in video tracking, it enables in determining the probability of definite and doubtful regions in cancer classification. Several examples and results would be provided to explain the aforesaid concepts. The talk concludes mentioning the challenging issues and the future directions of research including Big data analysis.
Agli studenti che parteciparanno al corso sarà attribuito un credito formativo universitario (CFU).
Per registrarsi all’evento