A Robust Multiple Feature Approach To Endpoint Detection In Car Environment Based On Advanced Classifiers

In this paper we propose an endpoint detection system based on the use of several features extracted from each speech frame, followed by a robust classifier (i.e Adaboost and Bagging of decision trees, and a multilayer perceptron) and a finite state automata (FSA). We present results for four different classifiers. The FSA module consisted of a 4-state decision logic that filtered false alarms and false positives. We compare the use of four different classifiers in this task. The look ahead of the method that we propose was of 7 frames, which are the number of frames that maximized the accuracy of the system. The system was tested with real signals recorded inside a car, with signal to noise ratio that ranged from 6 dB to 30dB. Finally we present experimental results demonstrating that the system yields robust endpoint detection.

Educational data mining and learning analytics : Clasificación de las matriculaciones de A.D.E. en la UOC

Trabajo de investigación que realiza un estudio clasificatorio de las asignaturas matriculadas en la carrera de Administración y Dirección de Empresas de la UOC en relación a su resultado. Se proponen diferentes métodos y modelos de comprensión del entorno en el que se realiza el estudio.

ICA as a preprocessing technique for classification

In this paper we propose the use of the independent component analysis (ICA) [1] technique for improving the classification rate of decision trees and multilayer perceptrons [2], [3]. The use of an ICA for the preprocessing stage, makes the structure of both classifiers simpler, and therefore improves the generalization properties. The hypothesis behind the proposed preprocessing is that an ICA analysis will transform the feature space into a space where the components are independent, and aligned to the axes and therefore will be more adapted to the way that a decision tree is constructed. Also the inference of the weights of a multilayer perceptron will be much easier because the gradient search in the weight space will follow independent trajectories. The result is that classifiers are less complex and on some databases the error rate is lower. This idea is also applicable to regression

Classifier combination for in vivo magnetic resonance spectra of brain tumours

In this paper we present a multi-stage classifier for magnetic resonance spectra of human brain tumours which is being developed as part of a decision support system for radiologists. The basic idea is to decompose a complex classification scheme into a sequence of classifiers, each specialising in different classes of tumours and trying to reproducepart of the WHO classification hierarchy. Each stage uses a particular set of classification features, which are selected using a combination of classical statistical analysis, splitting performance and previous knowledge.Classifiers with different behaviour are combined using a simple voting scheme in order to extract different error patterns: LDA, decision trees and the k-NN classifier. A special label named "unknown¿ is used when the outcomes of the different classifiers disagree. Cascading is alsoused to incorporate class distances computed using LDA into decision trees. Both cascading and voting are effective tools to improve classification accuracy. Experiments also show that it is possible to extract useful information from the classification process itself in order to helpusers (clinicians and radiologists) to make more accurate predictions and reduce the number of possible classification mistakes.

Stratification of the severity of critically ill patients with classification trees

Background: Development of three classification trees (CT) based on the CART (Classification and Regression Trees), CHAID (Chi-Square Automatic Interaction Detection) and C4.5 methodologies for the calculation of probability of hospital mortality; the comparison of the results with the APACHE II, SAPS II and MPM II-24 scores, and with a model based on multiple logistic regression (LR). Methods: Retrospective study of 2864 patients. Random partition (70:30) into a Development Set (DS) n = 1808 and Validation Set (VS) n = 808. Their properties of discrimination are compared with the ROC curve (AUC CI 95%), Percent of correct classification (PCC CI 95%); and the calibration with the Calibration Curve and the Standardized Mortality Ratio (SMR CI 95%). Results: CTs are produced with a different selection of variables and decision rules: CART (5 variables and 8 decision rules), CHAID (7 variables and 15 rules) and C4.5 (6 variables and 10 rules). The common variables were: inotropic therapy, Glasgow, age, (A-a)O2 gradient and antecedent of chronic illness. In VS: all the models achieved acceptable discrimination with AUC above 0.7. CT: CART (0.75(0.71-0.81)), CHAID (0.76(0.72-0.79)) and C4.5 (0.76(0.73-0.80)). PCC: CART (72(69- 75)), CHAID (72(69-75)) and C4.5 (76(73-79)). Calibration (SMR) better in the CT: CART (1.04(0.95-1.31)), CHAID (1.06(0.97-1.15) and C4.5 (1.08(0.98-1.16)). Conclusion: With different methodologies of CTs, trees are generated with different selection of variables and decision rules. The CTs are easy to interpret, and they stratify the risk of hospital mortality. The CTs should be taken into account for the classification of the prognosis of critically ill patients.

Evaluation of a revision of the BSPcast decision support system for control of brown spot of pear

Control of brown spot of pear requires fungicide treatments of pear trees during the growing season. Scheduling fungicide sprays with the Brown spot of pear forecasting system (BSPcast) provides significantfungicide savings but does not increase the efficacy of disease control. Modifications in BSPcast wereintroduced in order to increase system performance. The changes consisted of: (1) the use of a daily infectionrisk (Rm≥0.2) instead of the 3-day cumulative risk (CR≥0.4) to guide the fungicide scheduling, and (2) theinclusion of the effect of relative humidity during interrupted wetness periods. Trials were performed during2 years in an experimental pear orchard in Spain. The modifications introduced did not result in increaseddisease control efficacy, compared with the original BSPcast system. In one year, no reduction in the numberof fungicide applications was obtained using the modified BSPcast system in comparison to the original system, but in the second year the number of treatments was reduced from 15 to 13. The original BSPcast model overestimated the daily infection risk in 6.5% of days with wetness periods with low relative humidity during the wetness interruption, and in these cases the modified version was more adequate

Stable coalition structures with fixed decision scheme

This paper studies the stability of a finite local public goods economy in horizontal differentiation, where a jurisdiction's choice of the public good is given by an exogenous decision scheme. In this paper, we characterize the class of decision schemes that ensure the existence of an equilibrium with free mobility (that we call Tiebout equilibrium) for monotone distribution of players. This class contains all the decision schemes whose choice lies between the Rawlsian decision scheme and the median voter with mid-distance of the two median voters when there are ties. We show that for non-monotone distribution, there is no decision scheme that can ensure the stability of coalitions. In the last part of the paper, we prove the non-emptiness of the core of this coalition formation game

Restricted walks in regular trees

Let T be the Cayley graph of a finitely generated free group F. Given two vertices in T consider all the walks of a given length between these vertices that at a certain time must follow a number of predetermined steps. We give formulas for the number of such walks by expressing the problem in terms of equations in F and solving the corresponding equations.

Cooperative vs. non-cooperative mechanisms in the EU decision making process

According to the account of the European Union (EU) decision making proposed in this paper, this is a bargaining process during which actors shift their policy positions with a view to reaching agreements on controversial issues.

Generating trees for permutations avoiding generalized patterns

We construct generating trees with with one, two, and three labels for some classes of permutations avoiding generalized patterns of length 3 and 4. These trees are built by adding at each level an entry to the right end of the permutation, which allows us to incorporate the adjacency condition about some entries in an occurrence of a generalized pattern. We use these trees to find functional equations for the generating functions enumerating these classes of permutations with respect to different parameters. In several cases we solve them using the kernel method and some ideas of Bousquet-Mélou [2]. We obtain refinements of known enumerative results and find new ones.

Learning from Shocks and the Decision to Open

Two claims pervade the literature on the political economy of market reforms: that economic crises cause reforms; and that crises matter because they bring into question the validity of the economic model held to be responsible for them. Economic crises are said to spur a process of learning that is conducive to the abandonment of failing models and to the adoption of successful models. But although these claims have become the conventional wisdom, they have been hardly tested empirically due to the lack of agreement on what constitutes a crisis and to difficulties in measuring learning from them. I propose a model of rational learning from experience and apply it to the decision to open the economy. Using data from 1964 through 1990, I show that learning from the 1982 debt crisis was relevant to the first wave of adoption of an export promotion strategy, but learning was conditional on the high variability of economic outcomes in countries that opened up to trade. Learning was also symbolic in that the sheer number of other countries that liberalized was a more important driver of others’ decisions to follow suit.

The equation xpyq=zr and groups that act freely on Λ-trees

"Vegeu el resum a l’inici del document del fitxer adjunt."

Polynomial functors and trees

We explore the relationship between polynomial functors and trees. In the first part we characterise trees as certain polynomial functors and obtain a completely formal but at the same time conceptual and explicit construction of two categories of rooted trees, whose main properties we describe in terms of some factorisation systems. The second category is the category Ω of Moerdijk and Weiss. Although the constructions are motivated and explained in terms of polynomial functors, they all amount to elementary manipulations with finite sets. Included in Part 1 is also an explicit construction of the free monad on a polynomial endofunctor, given in terms of trees. In the second part we describe polynomial endofunctors and monads as structures built from trees, characterising the images of several nerve functors from polynomial endofunctors and monads into presheaves on categories of trees. Polynomial endofunctors and monads over a base are characterised by a sheaf condition on categories of decorated trees. In the absolute case, one further condition is needed, a projectivity condition, which serves also to characterise polynomial endofunctors and monads among (coloured) collections and operads.

A decision-making Fokker-Planck model in computational neuroscience

Minimal models for the explanation of decision-making in computational neuroscience are based on the analysis of the evolution for the average firing rates of two interacting neuron populations. While these models typically lead to multi-stable scenario for the basic derived dynamical systems, noise is an important feature of the model taking into account finite-size effects and robustness of the decisions. These stochastic dynamical systems can be analyzed by studying carefully their associated Fokker-Planck partial differential equation. In particular, we discuss the existence, positivity and uniqueness for the solution of the stationary equation, as well as for the time evolving problem. Moreover, we prove convergence of the solution to the the stationary state representing the probability distribution of finding the neuron families in each of the decision states characterized by their average firing rates. Finally, we propose a numerical scheme allowing for simulations performed on the Fokker-Planck equation which are in agreement with those obtained recently by a moment method applied to the stochastic differential system. Our approach leads to a more detailed analytical and numerical study of this decision-making model in computational neuroscience.

Level-wise node size distribution of randomly generated regular trees

"Vegeu el resum a l'inici del document del fitxer adjunt."