María Dolores Pérez-Godoy

CO2RBFN: an evolutionary cooperative---competitive RBFN design algorithm for classification problems

This paper presents a new evolutionary cooperative–competitive algorithm for the design of Radial Basis Function Networks (RBFNs) for classification problems. The algorithm, CO2RBFN, promotes a cooperative–competitive environment where each individual represents a radial basis function (RBF) and the entire population is responsible for the final solution. The proposal considers, in order to measure the credit assignment of an individual, three factors: contribution to the output of the complete RBFN, local error and overlapping. In addition, to decide the operators’ application probability over an RBF, the algorithm uses a Fuzzy Rule Based System. It must be highlighted that the evolutionary algorithm considers a distance measure which deals, without loss of information, with differences between nominal features which are very usual in classification problems. The precision and complexity of the network obtained by the algorithm are compared with those obtained by different soft computing methods through statistical tests. This study shows that CO2RBFN obtains RBFNs with an appropriate balance between accuracy and simplicity, outperforming the other methods considered.

Training algorithms for Radial Basis Function Networks to tackle learning processes with imbalanced data-sets

Graphical abstractDisplay Omitted HighlightsWe present a study about the performance of two classical weight training methods of Radial Basis Function Networks (RBFN), Least Mean Square (LMS) and Singular Value Decomposition (SVD), applied to classification problems, when the data-sets are imbalanced.These methods are tested with representative RBFN design paradigms: Clustering, Incremental, Genetic and CO2RBFN (a cooperative-competitive method proposed by the authors).The results obtained, statistically validated, show that SVD outperforms LMS, when the imbalance ratio of data-sets is low but when the imbalance ratio of these data sets grows, LMS outperforms SVD. Nowadays, many real applications comprise data-sets where the distribution of the classes is significantly different. These data-sets are commonly known as imbalanced data-sets. Traditional classifiers are not able to deal with these kinds of data-sets because they tend to classify only majority classes, obtaining poor results for minority classes. The approaches that have been proposed to address this problem can be categorized into three types: resampling methods, algorithmic adaptations and cost sensitive techniques.Radial Basis Function Networks (RBFNs), artificial neural networks composed of local models or RBFs, have demonstrated their efficiency in different machine learning areas. Centers, widths and output weights for the RBFs must be determined when designing RBFNs.Taking into account the locally tuned response of RBFs, the objective of this paper is to study the influence of global and local paradigms on the weights training phase, within the RBFNs design methodology, for imbalanced data-sets. Least Mean Square and the Singular Value Decomposition have been chosen as representatives of local and global weights training paradigms respectively. These learning algorithms are inserted into classical RBFN design methods that are run on imbalanced data-sets and also on these data-sets preprocessed with re-balance techniques. After applying statistical tests to the results obtained, some guidelines about the RBFN design methodology for imbalanced data-sets are provided.

MEFASD-BD

Nowadays, there is an incredible increase of data volumes around the world, with the Internet as one of the main actors in this scenario and a growth rate above 30GB/s. The treatment of this huge amount of information cannot be carried out through traditional data mining algorithms in an efficient way and it is necessary to adapt and design new algorithms towards distributed paradigms such as MapReduce. This situation is a challenge for the community, investigated under the widely known term of big data.This paper presents a new algorithm for the subgroup discovery task called MEFASD-BD. The algorithm is developed in Apache Spark based on the MapReduce paradigm, and it is able to tackle high dimensional datasets in an efficient way. In fact, this algorithm is the first approximation to big data within evolutionary fuzzy systems for subgroup discovery. MEFASD-BD implements novel MapReduce functions which are able to analyse the quality of the subgroups obtained for each map with respect to the original dataset in order to improve the quality of these subgroups. In addition, the final reduce function of the algorithm employs the token competition operator in order to select the best rules extracted in the different maps. An experimental study with high dimensional datasets is performed in order to show the advantages of this algorithm in this type of problems. Specifically, the results of the study show an important reduction of the runtime while keeping the values in the standard quality measures for subgroup discovery.

CO

This paper presents the adaptation of CO

^2

^2

RBFN for short-term forecasting of the extra virgin olive oil price in the Spanish market

RBFN, an evolutionary cooperative-competitive hybrid algorithm for the design of Radial Basis Function Networks, for short-term forecasting of the price of extra virgin olive oil. In the proposed cooperative-competitive environment, each individual represents a Radial Basis Function, and the entire population is responsible for the final solution. In order to calculate the application probability of the evolutive operators over a certain Radial Basis Function, a Fuzzy Rule Based System has been used. The olive oil time series have been analyzed using CO

CO 2 RBFN for Short and Medium Term Forecasting of the Extra-Virgin Olive Oil Price

^2

A study on the medium-term forecasting using exogenous variable selection of the extra-virgin olive oil with soft computing methods

RBFN. The results obtained have been compared with Auto-Regressive Integrated Moving Average (ARIMA) models and other data mining methods such as a fuzzy system developed with a GA-P algorithm, a multilayer perceptron trained with a conjugate gradient algorithm, and a radial basis function network trained using an LMS algorithm. The experimentation shows the high efficiency achieved by these methods, especially the data mining methods, which have slightly outperformed the ARIMA methodology.

An Study on Data Mining Methods for Short-Term Forecasting of the Extra Virgin Olive Oil Price in the Spanish Market

In this paper an adaptation of CO2RBFN, evolutionary COoperative- COmpetitive algorithm for Radial Basis Function Networks design, applied to the prediction of the extra-virgin olive oil price is presented. In this algorithm each individual represents a neuron or Radial Basis Function and the population, the whole network. Individuals compite for survival but must cooperate to built the definite solution. The forecasting of the extra-virgin olive oil price is addressed as a time series forecasting problem. In the experimentation medium-term predictions are obtained for first time with these data. Also short-term predictions with new data are calculated. The results of CO2RBFN have been compared with the traditional statistic forecasting Auto-Regressive Integrated Moving Average method and other data mining methods such as other neural networks models, a support vector machine method or a fuzzy system.

Comparative analysis of data mining and response surface methodology predictive models for enzymatic hydrolysis of pretreated olive tree biomass

Time series forecasting is an important task for the business sector. Agents involved in the olive oil sector consider that, for the olive oil price, medium-term predictions are more important than short-term predictions. In collaboration with these agents the forecasting of the price of extra-virgin olive oil six months ahead has been established as the aim of this work. According to expert opinion, the use of exogenous variables and technical indicators can help in this task and must be included in the forecasting process. The amount of variables that can be considered makes necessary the use of feature selection algorithms in order to reduce the number of variables and to increase the interpretability and usefulness of the obtained forecasting system. Thus, in this paper CO2RBFN, a cooperative-competitive algorithm for Radial Basis Function Network design, and other soft computing methods have been applied to the data sets with the whole set of input variables and to the data sets with the selected set of input variables. The experimentation carried out shows that CO2RBFN obtains the best results in medium term forecasting for olive oil prices with the whole and with the selected set of input variables. Moreover, the feature selection methods applied to the data sets highlighted some influential variables which could be considered not only for the prediction but also for the description of the complex process involved in the medium-term forecasting of the olive oil price.