10 research outputs found
Adaptive Basis Function Construction: An Approach for Adaptive Building of Sparse Polynomial Regression Models
No full textA comparison of heuristic methods for polynomial regression model induction
No full textWe compare four different heuristic methods for polynomial regression model induction. The methods are very different in their approaches. Our main concern in this study is in the differences of candidate model spaces the methods deal with (completely predefined versus non‐predefined), as well as search strategies used. We investigate the advantages and disadvantages of the approaches represented by the methods in terms of predictive error, complexity of the induced models and required computational resources. For empirical comparisons, we use twelve test problems.
First Published Online: 14 Oct 201
Adaptive Regression and Classification Models with Applications in Insurance
No full textNowadays, in the insurance industry the use of predictive modeling by means of regression and classification techniques is becoming increasingly important and popular. The success of an insurance company largely depends on the ability to perform such tasks as credibility estimation, determination of insurance premiums, estimation of probability of claim, detecting insurance fraud, managing insurance risk. This paper discusses regression and classification modeling for such types of prediction problems using the method of Adaptive Basis Function Constructio
Birzgalis 26
No full textJanis Birzgalis and friends in Edmonton, Alberta. Left to right: Peter Melnbarns, Janis Birzgalis, Edgars Vitols, Juris Cinats and son, John, Janis Jekabsons
Automatic extraction of geographic context from textual data
No full textThe amount of information on the internet grows exponentially. It isnot enough anymore just to have a general access to this huge amount of data,instead it is becoming a necessity to be able to use different kinds ofautomatic filters to retrieve just the information you actually want. Onesolution for the information filtering and retrieval is context analysis inwhich one of the contexts of interest is the geographic context. This paperstudies the problem and methodology of geoparsing – recognition of geographicnames in unstructured textual content for the aim of extracting geographiccontext. A prototype implementation of a geoparsing system, capable ofautomatically analyzing unstructured text, recognizing geographic informationand marking geographic names, is developed. Empirical evaluation of the systemusing articles from real-world news showed that the average quality of itsgeographic name recognition varies around 75-100%. Possible applications of thedeveloped prototype include automated grouping of any texts by their geographiccontexts (e.g., in news portals) and location-based search. Preliminary resultsof empirical evaluation showed that the average rate of its geographic namerecognition varies around 75-100%.DOI: http://dx.doi.org/10.15181/csat.v2i1.13</p
Interactive collaboration in distance learning based on remote control of server session
No full text
Robustness of empirical vibration correlation techniques for predicting the instability of unstiffened cylindrical composite shells in axial compression
No full textThin-walled carbon fiber reinforced plastic (CFRP) shells are increasingly used in aerospace industry. Such shells are prone to the loss of stability under compressive loads. Furthermore, the instability onset of monocoque shells exhibits a pronounced imperfection sensitivity. The vibration correlation technique (VCT) is being developed as a nondestructive test method for evaluation of the buckling load of the shells. In this study, accuracy and robustness of an existing and a modified VCT method are evaluated. With this aim, more than 20 thin-walled unstiffened CFRP shells have been produced and tested. The results obtained suggest that the vibration response under loads exceeding 0.25 of the linear buckling load needs to be characterized for a successful application of the VCT. Then the largest unconservative discrepancy of prediction by the modified VCT method amounted to ca. 22% of the critical load. Applying loads exceeding 0.9 of the buckling load reduced the average relative discrepancy to 6.4%.Aerospace Structures & Computational Mechanic
Does VDAC2 have a BH3 Domain For Binding Bax?
No full textCorresponding author (Biology): Lilian Ferkany, [email protected]://egrove.olemiss.edu/pharm_annual_posters_2022/1007/thumbnail.jp
