An architecture for making recommendations to courseware authors using association rule mining and collaborative filtering

Nowadays we find more and more applications for data mining techniques in e-learning and web-based adaptive educational systems. The useful information discovered can be used directly by the teacher or author of the course in order to improve instructional/learning performance. This can, however, imply a lot of work for the teacher who can greatly benefit from the help of educational recommender systems for doing this task. In this paper we propose a system oriented to find, share and suggest the most appropriate modifications to improve the effectiveness of the course. We describe an iterative methodology to develop and carry out the maintenance of web-based courses to which we have added a specific data mining step. We apply association rule mining to discover interesting information through students’ usage data in the form of IF-THEN recommendation rules. We have also used a collaborative recommender system to share and score the recommendation rules obtained by teachers with similar profiles along with other experts in education. Finally, we have carried out experiments with several real groups of students using a web-based adaptive course. The results obtained demonstrate that the proposed architecture constitutes a good starting point to future investigations in order to generalize the results over many course contents.     

A hybrid learning algorithm for a class of interval type-2 fuzzy neural networks

In real life, information about the world is uncertain and imprecise. The cause of this uncertainty is due to: deficiencies on given information, the fuzzy nature of our perception of events and objects, and on the limitations of the models we use to explain the world. The development of new methods for dealing with information with uncertainty is crucial for solving real life problems. In this paper three interval type-2 fuzzy neural network (IT2FNN) architectures are proposed, with hybrid learning algorithm techniques (gradient descent backpropagation and gradient descent with adaptive learning rate backpropagation). At the antecedents layer, a interval type-2 fuzzy neuron (IT2FN) model is used, and in case of the consequents layer an interval type-1 fuzzy neuron model (IT1FN), in order to fuzzify the rule’s antecedents and consequents of an interval type-2 Takagi-Sugeno-Kang fuzzy inference system (IT2-TSK-FIS). IT2-TSK-FIS is integrated in an adaptive neural network, in order to take advantage the best of both models. This provides a high order intuitive mechanism for representing imperfect information by means of use of fuzzy If-Then rules, in addition to handling uncertainty and imprecision. On the other hand, neural networks are highly adaptable, with learning and generalization capabilities. Experimental results are divided in two kinds: in the first one a non-linear identification problem for control systems is simulated, here a comparative analysis of learning architectures IT2FNN and ANFIS is done. For the second kind, a non-linear Mackey-Glass chaotic time series prediction problem with uncertainty sources is studied. Finally, IT2FNN proved to be more efficient mechanism for modeling real-world problems.     

Power series generalized nonlinear models

We introduce in this paper a new class of discrete generalized nonlinear models to extend the binomial, Poisson and negative binomial models to cope with count data. This class of models includes some important models such as log-nonlinear models, logit, probit and negative binomial nonlinear models, generalized Poisson and generalized negative binomial regression models, among other models, which enables the fitting of a wide range of models to count data. We derive an iterative process for fitting these models by maximum likelihood and discuss inference on the parameters. The usefulness of the new class of models is illustrated with an application to a real data set.     

A metaheuristic approach to the optimal definition of molecule-fixed axes in rovibrational Hamiltonians

We develop an algorithmic, metaheuristic approach to the definition of molecule-fixed axes orientation in molecules of arbitrary size. The goal is to simplify the treatment of overall rotation and rotation-vibration interaction in rovibrational Hamiltonians. Considering the kinetic elements of the rovibrational Hamiltonian, given by the G matrix, we select the optimal orientation of molecule-fixed axes minimizing specific G matrix elements. To such an end, we develop a global minimization method based in a hybrid Simulated Annealing+Powell's local minimization. The parameters of the method are calibrated using a set of non-rigid molecules: Acetaldehyde, glycolaldehyde, methyl formate and ethyl methyl ether. The results show that the principal axes of inertia do not give the simplest form to the pure rotational contribution. However, minimization of the G matrix rotational element does. Finally, we observe that in the cases considered it is not possible to nullify all the rotation-vibration coupling elements, since the torsional motions are coupled with the overall rotation. However, the treatment yields the optimal solution. The methodology proposed allows also to simplify simultaneously the pure rotational+rotation-vibration coupling elements in rovibrational Hamiltonians.     

LOCALIZING THE IRIS THROUGH MEMETIC ALGORITHMS

An iris recognition system requires efficient image processing techniques in order to duly represent and interpret the iris structural characteristics of an individual. The first processing stage should be the identification of the iris region in an eye image. This work introduces the application of evolutionary algorithms to localize the iris region in an eye image. A method based on memetic algorithms was proposed and used to find the circles that represent the external iris border and the pupil border. This method is applied after detecting the edges of the image through gradient algorithms. The efficiency of the memetic algorithm in solving the problem was compared to the application of the Wildes' method, which uses the Circular Hough Transform (CHT), a well-known algorithm employed to find circles in an edged image. To test the algorithms, images from the UBIRIS database (Proença and Alexandre 2005 Proença , H. and L. A. Alexandre . 2005 . UBIRIS: A noisy iris image database . ICIAP 2005, 13th Int. Conf. on Image Analysis and Processing , Cagliari, Ital. Lect. Notes Comput. Sci. 3617 : 970 – 977 . [Google Scholar]) were used.     

SEPARATE: a machine learning method based on semi-global partitions

Presents a machine learning method for solving classification and approximation problems. This method uses the divide-and-conquer algorithm design technique (taken from machine learning models based on a tree), with the aim of achieving design ease and good results on the training examples and allows semi-global actions on its computational elements (a feature taken from neural networks), with the aim of attaining good generalization and good behavior in the presence of noise in training examples. Finally, some results obtained after solving several problems with a particular implementation of SEPARATE are presented together with their analysis.     

Oral doxifluridine plus levoleucovorin in elderly patients with advanced breast cancer

BACKGROUND: There currently is no agreement regarding the appropriate treatment of elderly patients with advanced breast carcinoma (ABC). Doxifluridine (5-dFUR), a prodrug of 5-fluorouracil, has been found to be effective in this entity, but its use is limited by neurotoxicity and cardiotoxicity that are not observed when the oral formulation is used. The objective of this Phase II trial was to evaluate the effectiveness and tolerability of oral 5-dFUR, biomodulated with levoleucovorin (1-leucovorin), in elderly patients (age > 70 years) with ABC. METHODS: 5-dFUR was administered orally at 600 mg/m2 twice daily for 4 consecutive days every 12 days, and oral 1-leucovorin was administered as 25 mg 2 hours before each 5-dFUR administration. Response was assessed every five cycles according to the World Health Organization criteria. In the presence of response or stable disease, the patients were treated for a maximum of 15 cycles. RESULTS: Seventy-three eligible patients were enrolled, 27 of whom had been pretreated with chemotherapy and/or hormonotherapy; all were assessable for response and toxicity after a median follow-up of 15 months. The objective response rate was 26% (95% confidence interval, 17.4-45.4). Regression predominantly occurred in the presence of soft tissue involvement (skin, lymph nodes, and breast). The median time to response was 2 months (range, 1-2 months) and the median response duration was 7 months (range, 2-17+ months). The median survival was 24 months (range, 2-42+ months). The treatment was very well tolerated, and the side effects were manageable and always reversible. CONCLUSIONS: The results of the current study show that 5-dFUR plus 1-leucovorin, both given orally, are associated with excellent patient compliance. Although the results are suboptimal in terms of an objective response, this characteristic could allow 5-dFUR to be used in elderly patients considered unsuitable for "aggressive" chemotherapy.     

Voltammetric studies on mechanisms of dopamine efflux in the presence of substrates and cocaine from cells expressing human norepinephrine transporter

The effects of substrates m-tyramine and beta-phenethylamine, as well as cocaine, on the DA efflux from a cell line stably expressing the human norepinephrine transporter (hNET) were investigated by using rotating disk electrode voltammetry. Both the substrates and cocaine induced apparent DA efflux in a concentration-dependent manner. Their EC50 values for inducing DA efflux were similar to their IC50 values for inhibiting DA uptake. The substrate-induced DA efflux was inhibited by various NET blockers, enhanced by raising the internal [Na+] with Na+,K+-ATPase inhibition, but was insensitive to membrane potential-altering agents valinomycin, veratridine, and high [K+]. The initial rate of m-tyramine-induced DA efflux was related to preloaded [DA] in a manner defined by a Michaelis-Menten expression. In contrast, DA efflux in the presence of cocaine displayed a much slower efflux rate, lower efficacy, was not stimulated by elevated internal [Na+], and was nonsaturable with preloaded [DA]. Single exponential kinetic analysis of the entire time course of the DA efflux showed that the apparent first-order rate constant for m-tyramine-induced DA efflux declined with increased preloaded [DA], whereas that for the DA efflux in the presence of cocaine was unchanged with varying preloaded [DA]. These results suggest that the substrates stimulate the NET-dependent DA efflux by increasing the accessibility of the NET to internal DA, whereas cocaine "uncovers" NET-independent DA efflux by reducing the accessibility of diffused/leaked external DA to the NET.