Efficient training of RBF networks for classification

Radial Basis Function networks with linear outputs are often used in regression problems because they can be substantially faster to train than Multi-layer Perceptrons. For classification problems, the use of linear outputs is less appropriate as the outputs are not guaranteed to represent probabilities. In this paper we show how RBFs with logistic and softmax outputs can be trained efficiently using algorithms derived from Generalised Linear Models. This approach is compared with standard non-linear optimisation algorithms on a number of datasets.

Neural network regression with input uncertainty

It is generally assumed when using Bayesian inference methods for neural networks that the input data contains no noise or corruption. For real-world (errors in variable) problems this is clearly an unsafe assumption. This paper presents a Bayesian neural network framework which allows for input noise given that some model of the noise process exists. In the limit where this noise process is small and symmetric it is shown, using the Laplace approximation, that there is an additional term to the usual Bayesian error bar which depends on the variance of the input noise process. Further, by treating the true (noiseless) input as a hidden variable and sampling this jointly with the network's weights, using Markov Chain Monte Carlo methods, it is demonstrated that it is possible to infer the unbiassed regression over the noiseless input.

Bayesian inference for wind field retrieval

In many problems in spatial statistics it is necessary to infer a global problem solution by combining local models. A principled approach to this problem is to develop a global probabilistic model for the relationships between local variables and to use this as the prior in a Bayesian inference procedure. We show how a Gaussian process with hyper-parameters estimated from Numerical Weather Prediction Models yields meteorologically convincing wind fields. We use neural networks to make local estimates of wind vector probabilities. The resulting inference problem cannot be solved analytically, but Markov Chain Monte Carlo methods allow us to retrieve accurate wind fields.

Neural Network Modelling with Input Uncertainty: Theory and Application

It is generally assumed when using Bayesian inference methods for neural networks that the input data contains no noise. For real-world (errors in variable) problems this is clearly an unsafe assumption. This paper presents a Bayesian neural network framework which accounts for input noise provided that a model of the noise process exists. In the limit where the noise process is small and symmetric it is shown, using the Laplace approximation, that this method adds an extra term to the usual Bayesian error bar which depends on the variance of the input noise process. Further, by treating the true (noiseless) input as a hidden variable, and sampling this jointly with the network’s weights, using a Markov chain Monte Carlo method, it is demonstrated that it is possible to infer the regression over the noiseless input. This leads to the possibility of training an accurate model of a system using less accurate, or more uncertain, data. This is demonstrated on both the, synthetic, noisy sine wave problem and a real problem of inferring the forward model for a satellite radar backscatter system used to predict sea surface wind vectors.

Probabilistic methods for drug discovery

This thesis introduces a flexible visual data exploration framework which combines advanced projection algorithms from the machine learning domain with visual representation techniques developed in the information visualisation domain to help a user to explore and understand effectively large multi-dimensional datasets. The advantage of such a framework to other techniques currently available to the domain experts is that the user is directly involved in the data mining process and advanced machine learning algorithms are employed for better projection. A hierarchical visualisation model guided by a domain expert allows them to obtain an informed segmentation of the input space. Two other components of this thesis exploit properties of these principled probabilistic projection algorithms to develop a guided mixture of local experts algorithm which provides robust prediction and a model to estimate feature saliency simultaneously with the training of a projection algorithm.Local models are useful since a single global model cannot capture the full variability of a heterogeneous data space such as the chemical space. Probabilistic hierarchical visualisation techniques provide an effective soft segmentation of an input space by a visualisation hierarchy whose leaf nodes represent different regions of the input space. We use this soft segmentation to develop a guided mixture of local experts (GME) algorithm which is appropriate for the heterogeneous datasets found in chemoinformatics problems. Moreover, in this approach the domain experts are more involved in the model development process which is suitable for an intuition and domain knowledge driven task such as drug discovery. We also derive a generative topographic mapping (GTM) based data visualisation approach which estimates feature saliency simultaneously with the training of a visualisation model.

From the physics of interacting polymers to optimizing routes on the London Underground

Optimizing paths on networks is crucial for many applications, ranging from subway traffic to Internet communication. Because global path optimization that takes account of all path choices simultaneously is computationally hard, most existing routing algorithms optimize paths individually, thus providing suboptimal solutions. We use the physics of interacting polymers and disordered systems to analyze macroscopic properties of generic path optimization problems and derive a simple, principled, generic, and distributed routing algorithm capable of considering all individual path choices simultaneously. We demonstrate the efficacy of the algorithm by applying it to: (i) random graphs resembling Internet overlay networks, (ii) travel on the London Underground network based on Oyster card data, and (iii ) the global airport network. Analytically derived macroscopic properties give rise to insightful new routing phenomena, including phase transitions and scaling laws, that facilitate better understanding of the appropriate operational regimes and their limitations, which are difficult to obtain otherwise.

Unsupervised KDD to creatively support managers' decision making with fuzzy association rules:a distribution channel application

To be competitive in contemporary turbulent environments, firms must be capable of processing huge amounts of information, and effectively convert it into actionable knowledge. This is particularly the case in the marketing context, where problems are also usually highly complex, unstructured and ill-defined. In recent years, the development of marketing management support systems has paralleled this evolution in informational problems faced by managers, leading to a growth in the study (and use) of artificial intelligence and soft computing methodologies. Here, we present and implement a novel intelligent system that incorporates fuzzy logic and genetic algorithms to operate in an unsupervised manner. This approach allows the discovery of interesting association rules, which can be linguistically interpreted, in large scale databases (KDD or Knowledge Discovery in Databases.) We then demonstrate its application to a distribution channel problem. It is shown how the proposed system is able to return a number of novel and potentially-interesting associations among variables. Thus, it is argued that our method has significant potential to improve the analysis of marketing and business databases in practice, especially in non-programmed decisional scenarios, as well as to assist scholarly researchers in their exploratory analysis. © 2013 Elsevier Inc.

Implementation of Genetic Algorithms for Transit Points Arrangement

The problem of transit points arrangement is presented in the paper. This issue is connected with accuracy of tariff distance calculation and it is the urgent problem at present. Was showed that standard method of tariff distance discovering is not optimal. The Genetic Algorithms are used in optimization problem resolution. The UML application class diagram and class content are showed. In the end the example of transit points arrangement is represented.

Comparative Analysis of Viscoelastic Models Involving Fractional Derivatives of Different Orders

Mathematics Subject Classification: 74D05, 26A33

Fractional Integration and Fractional Differentiation of the M-Series

Mathematics Subject Classification: 26A33, 33C60, 44A15

Codifferentiable Mappings with Applications to Vector Optimality

AMS subject classification: Primary 49J52; secondary: 26A27, 90C48, 47N10.

An inconsistency control system based on incomplete pairwise comparison matrices

Incomplete pairwise comparison matrix was introduced by Harker in 1987 for the case in which the decision maker does not fill in the whole matrix completely due to, e.g., time limitations. However, incomplete matrices occur in a natural way even if the decision maker provides a completely filled in matrix in the end. In each step of the total n(n–1)/2, an incomplete pairwise comparison is given, except for the last one where the matrix turns into complete. Recent results on incomplete matrices make it possible to estimate inconsistency indices CR and CM by the computation of tight lower bounds in each step of the filling in process. Additional information on ordinal inconsistency is also provided. Results can be applied in any decision support system based on pairwise comparison matrices. The decision maker gets an immediate feedback in case of mistypes, possibly causing a high level of inconsistency.

Néhány megjegyzés a kockázat, bizonytalanság, valószínűség kérdéséhez (Some remarks about uncertainty, risk and probability)

A címben említett három fogalom a közgazdasági elméletben központi szerepet foglal el. Ezek viszonya elsősorban a közgazdaságtudományi megismerés határait feszegeti. Mit tudunk a gazdasági döntésekről? Milyen információk alapján születnek a döntések? Lehet-e a gazdasági döntéseket „tudományos” alapra helyezni? A bizonytalanság kérdéséről az 1920-as években való megjelenése óta mindent elmondtak. Megvizsgálták a kérdést filozófiailag, matematikailag. Tárgyalták a kérdés számtalan elméleti és gyakorlati aspektusát. Akkor miért kell sokadszorra is foglalkozni a témával? A válasz igen egyszerű: azért, mert a kérdés minden szempontból ténylegesen alapvető, és mindenkor releváns. Úgy hírlik, hogy a római diadalmenetekben a győztes szekerén mindig volt egy rabszolga is, aki folyamatosan figyelmeztette a diadaltól megmámorosodott vezért, hogy ő is csak egy ember, ezt ne feledje el. A gazdasági döntéshozókat hasonló módon újra és újra figyelmeztetni kell arra, hogy a gazdasági döntések a bizonytalanság jegyében születnek. A gazdasági folyamatok megérthetőségének és kontrollálhatóságának van egy igen szoros korlátja. Ezt a korlátot a folyamatok inherens bizonytalansága adja. A gazdasági döntéshozók fülébe folyamatosan duruzsolni kell: ők is csak emberek, és ezért ismereteik igen korlátozottak. A „bátor” döntések során az eredmény bizonytalan, a tévedés azonban bizonyosra vehető. / === / In the article the author presents some remarks on the application of probability theory in financial decision making. From mathematical point of view the risk neutral measures used in finance are some version of separating hyperplanes used in optimization theory and in general equilibrium theory. Therefore they are just formally a probabilities. They interpretation as probabilities are misleading analogies leading to wrong decisions.

Improving the predictability of time series by chaotic parameter estimation

Limited literature regarding parameter estimation of dynamic systems has been identified as the central-most reason for not having parametric bounds in chaotic time series. However, literature suggests that a chaotic system displays a sensitive dependence on initial conditions, and our study reveals that the behavior of chaotic system: is also sensitive to changes in parameter values. Therefore, parameter estimation technique could make it possible to establish parametric bounds on a nonlinear dynamic system underlying a given time series, which in turn can improve predictability. By extracting the relationship between parametric bounds and predictability, we implemented chaos-based models for improving prediction in time series. ^ This study describes work done to establish bounds on a set of unknown parameters. Our research results reveal that by establishing parametric bounds, it is possible to improve the predictability of any time series, although the dynamics or the mathematical model of that series is not known apriori. In our attempt to improve the predictability of various time series, we have established the bounds for a set of unknown parameters. These are: (i) the embedding dimension to unfold a set of observation in the phase space, (ii) the time delay to use for a series, (iii) the number of neighborhood points to use for avoiding detection of false neighborhood and, (iv) the local polynomial to build numerical interpolation functions from one region to another. Using these bounds, we are able to get better predictability in chaotic time series than previously reported. In addition, the developments of this dissertation can establish a theoretical framework to investigate predictability in time series from the system-dynamics point of view. ^ In closing, our procedure significantly reduces the computer resource usage, as the search method is refined and efficient. Finally, the uniqueness of our method lies in its ability to extract chaotic dynamics inherent in non-linear time series by observing its values. ^