Technical Reports
http://hdl.handle.net/11141/2867
2019-09-08T22:08:30ZA GIS-BASED SYNTHESIS OF RELICT SHORELINES IN PENINSULAR FLORIDA
http://hdl.handle.net/11141/2810
A GIS-BASED SYNTHESIS OF RELICT SHORELINES IN PENINSULAR FLORIDA
Hardin, Justin O.
A literature review as well as a consultation with a geomorphologist, Dr. Christopher P.
Williams, at the Florida Geological Survey (FGS) were both performed in order to understand
the state of knowledge on relict shorelines in peninsular Florida. This knowledge was compiled
into a map, Figure 1, that shows the locations and relative ages of geologically confirmed relict
shorelines in the Florida peninsula. A number of additional required areas of research in order to
“complete” the map are then discussed. The project highlights both the rapidity of Florida’s
Pliocene-Pleistocene coastal evolution and the relatively high amount of incompleteness still
present in the current state of knowledge on Florida relict shorelines.
2019-04-01T00:00:00ZLearning implicit user interest hierarchy for context in personalization
http://hdl.handle.net/11141/842
Learning implicit user interest hierarchy for context in personalization
Kim, Hyoung-Rae; Chan, Philip K.
To provide a more robust context for personalization, we desire to extract a continuum of general (long-term) to specific (short-term) interests of a user. Our proposed approach is to learn a user interest hierarchy (UIH) from a set of web pages visited by a user. We devise a divisive hierarchical clustering (DHC) algorithm to group words (topics) into a hierarchy where more general interests are represented by a larger set of words. Each web page can then be assigned to nodes in the hierarchy for further processing in learning and predicting interests. This approach is analogous to building a subject taxonomy for a library catalog system and assigning books to the taxonomy. Our approach does not need user involvement and learns the UIH "implicitly." Furthermore, it allows the original objects, web pages, to be assigned to multiple topics (nodes in the hierarchy). In this paper, we focus on learning the UIH from a set of visited pages. We propose a few similarity functions and dynamic threshold-finding methods, and evaluate the resulting hierarchies according to their meaningfulness and shape.
2002-10-05T00:00:00ZAn algorithm applicable to clearing combinatorial exchanges
http://hdl.handle.net/11141/841
An algorithm applicable to clearing combinatorial exchanges
Silaghi, Marius-Calin
It is important to approach negotiations in a way that ensures privacy. So far, research has focused on securely solving restricted classes of negotiation techniques, mainly the (M+1)-st-price auctions. Here we show how these results can be adapted to more general problems. This paper extends our previous results on how distributed finite discrete problems can be solved securely. Such problems can model larger classes of negotiation problems, .e.g. Combinatorial Exchanges [Sil02]. In Finite Discrete Maximization, each tuple in the problem space is associated with an integer value in a predefined interval and we search for a maximizing input. Values from different subproblems are combined additively. We show that unconstrained distributed Finite Discrete Maximization problems can be solved securely using a scheme that we propose for translating shared secret values into shared differential bids. Differential bid vectors are already used in [AS02][Bra02]. Constrained distributed Finite Discrete Maximization poses additional challenges, due to the loss of additivity of the maximized cost, when infeasibility is marked as the lowest finite value. We found two ways of solving this problem: a) by using an additional multiplication value; and b) by using larger variable domains. While the first alternative enforces a threshold to the privacy level in our current protocol, the second one increases much the complexity of the computation. The proposed algorithms are only (t/3)-private, where t is the number of participants.
2002-09-16T00:00:00ZMinimizing N-points interpolation curvature, heuristics for solutions using arcs and lines
http://hdl.handle.net/11141/177
Minimizing N-points interpolation curvature, heuristics for solutions using arcs and lines
Vishen, Rahul; Silaghi, Marius Cӑlin
Knowing a set of points on a curve, the interpolation problem is to hypothesize the location of the intermediary ones. A large set of interpolation techniques are known. We address the problem of generating a path with minimal maximum curvature, passing through N ordered points and joining the two end-points at predefined directions. This is related to R-geodesics, which have been used to generate paths with minimum average curvature between two given points that have to be joined at predefined directions and curvature. For example, when interpolating GPS points to reconstruct a vehicle’s trajectory, we may know that the centripetal acceleration is upper bounded due to physical constraints, hence adding constraints on the trajectory curvature. Among two interpolations with the same maximum curvature, we prefer the one with shorter trajectory. We compare experimentally several interpolations techniques, and propose heuristics to generate paths based on concatenated arc and line segments (also known as R-geodesics) inferred based on tuples of three
consecutive points. Benchmarks with over 1000 simulated and real scenarios show that this algorithm is 73% percent better then the next candidate method we propose and which is based on bi-arcs with hillclimbing. A remaining open question is whether a global optima can be achieved and proven.
2013-02-11T00:00:00Z