Past Events
Primary tabs
MSc Seminar: "Linking Historical Censuses to generate Longitudinal data using discrete methods" by Pavneet Kaur
ABSTRACT:
In Data Mining, Record Linkage is the process of identifying the same entities in one or more data sources in the absence of unique identifiers. Social Scientists and Historians have employed various linkage techniques to construct the longitudinal data by linking two or more historical censuses. Longitudinal data can provide us with valuable information about characteristics of population change over time. However, construction of such longitudinal data is challenged by the unavailability of personal identifiers.
PhD Seminar: Stefan Schneider
ABSTRACT:
PhD Defence: Jamileh Yousefi - "A Neuro-fuzzy Classifier for Datasets with Skewed Feature Values"
ABSTRACT:
Most machine learning algorithms perform poorly on skewed datasets. Data distributions in machine learning, when they are discussed at all, are generally expected to have a symmetric distribution, if not actually be normally distributed. The impact of skewed data distribution on the performance of machine learning algorithms has not been given much attention.