[Legacy Report] (Day 2) 5th International Conference on Fuzzy and Neural Computing (FANCCO 2015) : vTools Events

IEEE.org | IEEE Xplore Digital Library | IEEE Standards | IEEE Spectrum | More Sites

[Legacy Report] (Day 2) 5th International Conference on Fuzzy and Neural Computing (FANCCO 2015)

#LUBE #Method #Smart #grids #graph #matching #fuzzy #techniques. #Big #data

Key note 2 : Uncertainty Handling using Neural Network-Based Prediction Intervals

Prof. Dipti Srinivasan,
NUS, Singapore

Abstract : Prediction Intervals (PI) are powerful tools to quantify the potential uncertainties associated with datasets, and are often used in conjunction with neural networks for forecasting applications. Prediction intervals are comprised of upper and lower bounds that envelope the unknown values with a probability termed as a confidence level. The primary objective for constructing the PIs is to improve the prediction accuracy by quantifying the uncertainties associated with point forecasts. These PIs can then be efficiently used in the decision making process and for evaluation of multiple scenarios that the decision makers often face in large scale practical problems. This talk will discuss prediction-interval based uncertainty modeling methodologies for decision making and risk assessment, and its application in a smart grid environment.

Key Note 3: A Fuzzy Approach to Big Data Analytics Problems

Dr. Raghuram Krishnapuram ,
Program Manager, Financial Services, Xerox Innovation Group, India

Abstract:

With the explosion of information sources, sensor data and social media, unstructured and noisy data has become ubiquitous. In the coming years of "big data," the proportion of uncertain data will only increase. The information management and data analytics communities need to find new ways of addressing this problem by devising techniques to deal with the four v's of big data, namely volume, variety, velocity and veracity. This talk will describe why fuzzy techniques are more relevant now than ever before, in order to deal with this challenge. In particular, the talk will address issues related to the "four v's" in clustering as well as other combinatorial optimization problems such as graph matching, and also outline recent developments in soft representations that have made a significant difference in the areas of natural language processing and text analytics.