Search

The field of data mining is dedicated to the analysis of data in order to find underlying connections and the discovery of new patterns. Since the volume of data to be analyzed is sometimes quite significant, there is the need for efficient data mining algorithms to be implemented. The market-basket algorithm can represent a breakthrough in data mining techniques. As the associations that are to be analyzed grow more and more abstract, the market-basket approach is unable to deal with imprecise temporal associations, leaving a big area uncharted. This research is dedicated to the analysis of temporal imprecise associations through the modification of a standard a-priori approach by means of fuzzy set relations to classify the associations relating different sources of data. The results of this research show that it is possible to investigate such relations with the help of fuzzy set classification for temporal associations, and the result of such exploration is as easily understandable as the standard a-priori algorithm.

Some of the algorithms developed within the artificial neural-networks tradition can be easily adopted to wireless sensor network platforms and in the same time they can meet the requirements for sensor networks like: simple parallel distributed computation, distributed storage, data robustness and auto-classification of sensor readings. Dimensionality reduction, obtained simply from the outputs of the neural-networks clustering algorithms, leads to lower communication costs and energy savings. Two different data aggregation architectures are presented, with algorithms which use wavelets for initial data-processing of the sensory inputs and artificial neural-networks which use unsupervised learning methods for categorization of the sensory inputs. They are analyzed on a data obtained from a set of several motes, equipped with several sensors each. Results from deliberately simulated malfunctioning sensors show the data robustness of these architectures.

In this paper we model the conjecture that stress has a negative impact on imitation-based learning, in the context of the IETAL and MASIVE theories. We pose a central question: Which is worse - a loss of available information due to stress, or the acquisition of misinformation due to stress? This question is addressed by modeling a modified IETAL agent in the MASIVE environment for which imitation-based learning is affected by stress. We present an algebraic representation of the modified agent, thus giving an extension of earlier work designed to explore learning phenomena in human subjects. Our goal is to break ground in a new direction in order to develop a richer theory based on our previous work in the domain of cognitive robotics and multiagent systems.

This paper overviews the undergraduate research and teaching activities at the Cognitive Agency and Robotics Laboratory (CARoL) at Towson University. The lab emerged out of an experiment conducted in a small classroom setting. Today, it can be used as a recipe for building a setting where quality undergraduate (and graduate) research projects in robotics can be conducted on a shoestring.