Gianluigi Folino

An Adaptive Distributed Ensemble Approach to Mine Concept-Drifting Data Streams

An adaptive boosting ensemble algorithm for classifying homogeneous distributed data streams is presented. The method builds an ensemble of classifiers by using Genetic Programming (GP) to inductively generate decision trees, each trained on different parts of the distributed training set. The approach adopts a co-evolutionary platform to support a cooperative model of GP. A change detection strategy, based on self-similarity of the ensemble behavior, and measured by its fractal dimension, permits to capture time- evolving trends and patterns in the stream, and to reveal changes in evolving data streams. The approach tracks online ensemble accuracy deviation over time and decides to recompute the ensemble if the deviation has exceeded a pre- specified threshold. This allows the maintenance of an accurate and up-to-date ensemble of classifiers for continuous flows of data with concept drifts. Experimental results on a real life data set show the validity of the approach.

Genetic Programming and Simulated Annealing: A Hybrid Method to Evolve Decision Trees

A method for the data mining task of data classification, suitable to be implemented on massively parallel architectures, is proposed. The method combines genetic programming and simulated annealing to evolve a population of decision trees. A cellular automaton is used to realise a fine-grained parallel implementation of genetic programming through the diffusion model and the annealing schedule to decide the acceptance of a new solution. Preliminary experimental results, obtained by simulating the behaviour of the cellular automaton on a sequential machine, show significant better performances with respect to C4.5.

A scalable cellular implementation of parallel genetic programming

A new parallel implementation of genetic programming (GP) based on the cellular model is presented and compared with both canonical GP and the island model approach. The method adopts a load-balancing policy that avoids the unequal utilization of the processors. Experimental results on benchmark problems of different complexity show the superiority of the cellular approach with respect to the canonical sequential implementation and the island model. A theoretical performance analysis reveals the high scalability of the implementation realized and allows to predict the size of the population when the number of processors and their efficiency are fixed.

GP ensembles for large-scale data classification

An extension of cellular genetic programming for data classification (CGPC) to induce an ensemble of predictors is presented. Two algorithms implementing the bagging and boosting techniques are described and compared with CGPC. The approach is able to deal with large data sets that do not fit in main memory since each classifier is trained on a subset of the overall training data. The predictors are then combined to classify new tuples. Experiments on several data sets show that, by using a training set of reduced size, better classification accuracy can be obtained, but at a much lower computational cost

Parallel hybrid method for SAT that couples genetic algorithms and local search

A parallel hybrid method for solving the satisfiability (SAT) problem that combines cellular genetic algorithms (GAs) and the random walk SAT (WSAT) strategy of greedy SAT (GSAT) is presented. The method, called cellular genetic WSAT (CGWSAT), uses a cellular GA to perform a global search from a random initial population of candidate solutions and a local selective generation of new strings. The global search is then specialized in local search by adopting the WSAT strategy. A main characteristic of the method is that it indirectly provides a parallel implementation of WSAT when the probability of crossover is set to zero. CGWSAT has been implemented on a Meiko CS-2 parallel machine using a 2D cellular automaton as a parallel computation model. The algorithm has been tested on randomly generated problems and some classes of problems from the DIMACS and SATLIB test set.

GP ensemble for distributed intrusion detection systems

In this paper an intrusion detection algorithm based on GP ensembles is proposed. The algorithm runs on a distributed hybrid multi-island model-based environment to monitor security-related activity within a network. Each island contains a cellular genetic program whose aim is to generate a decision-tree predictor, trained on the local data stored in the node. Every genetic program operates cooperatively, yet independently by the others, by taking advantage of the cellular model to exchange the outmost individuals of the population. After the classifiers are computed, they are collected to form the GP ensemble. Experiments on the KDD Cup 1999 Data show the validity of the approach.

An adaptive flocking algorithm for performing approximate clustering

This paper presents an approach based on an adaptive bio-inspired method to make state of the art clustering algorithms scalable and to provide them with an any-time behavior. The method is based on the biology-inspired paradigm of a flock of birds, i.e. a population of simple agents interacting locally with each other and with the environment. The flocking algorithm provides a model of decentralized adaptive organization useful to solve complex optimization, classification and distributed control problems. This approach avoids the sequential search of canonical clustering algorithms and permits a scalable implementation.The method is applied to design two novel clustering algorithms based on the main principles of two popular clustering algorithms: DBSCAN and SNN. This apporach can identify clusters of widely varying shapes and densities and is able to extract an approximate view of the clusters whenever it is required. Both the algorithms have been evaluated on synthetic and real world data sets and the impact of the flocking strategy on performance has been evaluated.

Distributed Nearest Neighbor-Based Condensation of Very Large Data Sets

In this work, the parallel fast condensed nearest neighbor (PFCNN) rule, a distributed method for computing a consistent subset of a very large data set for the nearest neighbor classification rule is presented. In order to cope with the communication overhead typical of distributed environments and to reduce memory requirements, different variants of the basic PFCNN method are introduced. An analysis of spatial cost, CPU cost, and communication overhead is accomplished for all the algorithms. Experimental results, performed on both synthetic and real very large data sets, revealed that these methods can be profitably applied to enormous collections of data. Indeed, they scale up well and are efficient in memory consumption, confirming the theoretical analysis, and achieve noticeable data reduction and good classification accuracy. To the best of our knowledge, this is the first distributed algorithm for computing a training set consistent subset for the nearest neighbor rule.

Combining cellular genetic algorithms and local search for solving satisfiability problems

A new parallel hybrid method for solving the satisfiability problem that combines cellular genetic algorithms and the random walk (WSAT) strategy of GSAT is presented. The method, called CGWSAT, uses a cellular genetic algorithm to perform a global search on a random initial population of candidate solutions and a local selective generation of new strings. Global search is specialized in local search by adopting the WSAT strategy. CGWSAT has been implemented on a Meiko CS-2 parallel machine using a two-dimensional cellular automaton as a parallel computation model. The algorithm has been tested on randomly generated problems and some classes of problems from the DIMACS test set.

A model based on cellular automata for the parallel simulation of 3D unsaturated flow

Cellular automata (CA) are discrete dynamic systems that are used for modeling many physical systems. They are often used as an alternative to model and solve large-scale systems where the use of partial differential equations involve complex and computationally expensive simulations. The purpose of this work is to investigate the use of CA based techniques for modeling and parallel simulation of water flux in unsaturated soils. Unsaturated flow processes are an important topic in several branches of hydrology, soil science and agricultural engineering dealing with soil-atmosphere interaction, subsurface flow and transport processes. In this paper a CA model for 3D unsaturated flow simulation is proposed using an extension of the original computational paradigm of cellular automata. This model, aimed at simulating large-scale systems, uses a macroscopic CA approach where local laws with a clear physical meaning govern interactions among automata. Its correctness is proved by CAMELot system, which allows the specification, parallel simulation, visualization, steering and analysis of CA models in the same environment, using a friendly interface and providing at the same time considerable flexibility. The model has been validated with reference multidimensional solutions taken from benchmarks in literature, showing a good agreement, even in the cases where non-linearity was very marked. Furthermore, using some of these benchmarks we present a scalability analysis of the model and different quantization techniques aimed at reducing the number of messages exchanged and the execution time when simulations are characterized by scarce mass interactions.