Min-Thai Wu

An effective parallel approach for genetic-fuzzy data mining

Data mining is most commonly used in attempts to induce association rules from transaction data. In the past, we used the fuzzy and GA concepts to discover both useful fuzzy association rules and suitable membership functions from quantitative values. The evaluation for fitness values was, however, quite time-consuming. Due to dramatic increases in available computing power and concomitant decreases in computing costs over the last decade, learning or mining by applying parallel processing techniques has become a feasible way to overcome the slow-learning problem. In this paper, we thus propose a parallel genetic-fuzzy mining algorithm based on the master-slave architecture to extract both association rules and membership functions from quantitative transactions. The master processor uses a single population as a simple genetic algorithm does, and distributes the tasks of fitness evaluation to slave processors. The evolutionary processes, such as crossover, mutation and production are performed by the master processor. It is very natural and efficient to run the proposed algorithm on the master-slave architecture. The time complexities for both sequential and parallel genetic-fuzzy mining algorithms have also been analyzed, with results showing the good effect of the proposed one. When the number of generations is large, the speed-up can be nearly linear. The experimental results also show this point. Applying the master-slave parallel architecture to speed up the genetic-fuzzy data mining algorithm is thus a feasible way to overcome the low-speed fitness evaluation problem of the original algorithm.

An ACS-based framework for fuzzy data mining

Data mining is often used to find out interesting and meaningful patterns from huge databases. It may generate different kinds of knowledge such as classification rules, clusters, association rules, and among others. A lot of researches have been proposed about data mining and most of them focused on mining from binary-valued data. Fuzzy data mining was thus proposed to discover fuzzy knowledge from linguistic or quantitative data. Recently, ant colony systems (ACS) have been successfully applied to optimization problems. However, few works have been done on applying ACS to fuzzy data mining. This thesis thus attempts to propose an ACS-based framework for fuzzy data mining. In the framework, the membership functions are first encoded into binary-bits and then fed into the ACS to search for the optimal set of membership functions. The problem is then transformed into a multi-stage graph, with each route representing a possible set of membership functions. When the termination condition is reached, the best membership function set (with the highest fitness value) can then be used to mine fuzzy association rules from a database. At last, experiments are made to make a comparison with other approaches and show the performance of the proposed framework.

A multi-level ant-colony mining algorithm for membership functions

Fuzzy data mining is used to extract fuzzy knowledge from linguistic or quantitative data. It is an extension of traditional data mining and the derived knowledge is relatively meaningful to human beings. In the past, we proposed a mining algorithm to find suitable membership functions for fuzzy association rules based on ant colony systems. In that approach, precision was limited by the use of binary bits to encode the membership functions. This paper elaborates on the original approach to increase the accuracy of results by adding multi-level processing. A multi-level ant colony framework is thus designed and an algorithm based on the structure is proposed to achieve the purpose. The proposed approach first transforms the fuzzy mining problem into a multi-stage graph, with each route representing a possible set of membership functions. The new approach then extends the previous one, using multi-level processing to solve the problem in which the maximum quantities of item values in the transactions may be large. The membership functions derived in a given level will be refined in the subsequent level. The final membership functions in the last level are then outputted to the rule-mining phase to find fuzzy association rules. Experiments are also performed to show the performance of the proposed approach. The experimental results show that the proposed multi-level ant colony systems mining approach can obtain improved results.

A continuous ant colony system framework for fuzzy data mining

The goal of data mining is to find out interesting and meaningful patterns from large databases. In some real applications, many data are quantitative and linguistic. Fuzzy data mining was thus proposed to discover fuzzy knowledge from this kind of data. In the past, two mining algorithms based on the ant colony systems were proposed to find suitable membership functions for fuzzy association rules. They transformed the problem into a multi-stage graph, with each route representing a possible set of membership functions, and then, used the any colony system to solve it. They, however, searched for solutions in a discrete solution space in which the end points of membership functions could be adjusted only in a discrete way. The paper, thus, extends the original approaches to continuous search space, and a fuzzy mining algorithm based on the continuous ant approach is proposed. The end points of the membership functions may be moved in the continuous real-number space. The encoding representation and the operators are also designed for being suitable in the continuous space, such that the actual global optimal solution is contained in the search space. Besides, the proposed approach does not have fixed edges and nodes in the search process. It can dynamically produce search edges according to the distribution functions of pheromones in the solution space. Thus, it can get a better nearly global optimal solution than the previous two ant-based fuzzy mining approaches. The experimental results show the good performance of the proposed approach as well.

Extracting membership functions in fuzzy data mining by Ant Colony Systems

Ant colony systems (ACS) have been successfully applied to optimization problems in recent years. However, few works have been done on applying ACS to data mining. This paper proposes an ACS-based algorithm to extract membership functions in fuzzy data mining. The membership functions are first encoded into binary bits and then fed into the ACS to search for the optimal set of membership functions. An example is given to demonstrate the proposed algorithm. Numerical experiments are also made to show the performance of the proposed approach.

A Hierarchical Gene-Set Genetic Algorithm

In this paper, gene sets, instead of individual genes, are used in the genetic process to speed up convergence. A gene-set mutation operator is proposed, which can make several neighboring genes to simultaneously mutate. A gene-set crossover operator is also designed to choose the crossover points at the boundary of gene sets. The proposed gene-set mutation and crossover operators will cause a larger diversity than the conventional ones. A hierarchical gene-set genetic algorithm is then proposed, which uses adjustable gene-set lengths to find final solutions. Different phases of populations use different gene-set lengths to perform the genetic operations. The gene-set length is shortened in half in each phase until the length is 1. Experiments on three problems are also made to show the effectiveness of the proposed gene-set genetic algorithm.

Using the master-slave parallel architecture for genetic-fuzzy data mining

Data mining is most commonly used in attempts to induce association rules from transaction data. In the past, we used the fuzzy and GA concepts to discover both useful fuzzy association rules and suitable membership functions from quantitative values. The evaluation for fitness values was, however, quite time-consuming. In this paper, we thus propose a parallel genetic-fuzzy mining algorithm based on the master-slave architecture to extract both association rules and membership functions from quantitative transactions. The master processor uses a single population as a simple genetic algorithm does, and distributes the tasks of fitness evaluation to slave processors. The evolutionary processes, such as crossover, mutation and production are performed by the master processor. Both the theoretic analysis and the experimental results show that the speed-up of the proposed parallel algorithm can increase nearly linear along with the number of individuals to be evaluated.

Using dynamic mutation rates in gene-set genetic algorithms

In this paper, we use dynamic mutation rates in gene-set genetic algorithms to increase the number of offspring when the gene-set size is large. Experiments on three problems are made to show the effectiveness of the genetic algorithm with dynamic mutation rates. From the experimental results, the proposed algorithm can get better convergence effects than the one with fixed mutation rates and than the simple GA, but spends only a little more computational time.

Using escape operations in gene-set genetic algorithms

In the past, gene-set genetic algorithms were proposed, in which gene sets, instead of individual genes, were used in the genetic process to speed up the convergence. In this paper, another escape operation, as well as the mutation operation, is designed for gene-set genetic algorithms to increase the probability of finding global optima. The property that a longer gene set will cause a larger diversity is shown. An escape operation based on the property is thus designed and a modified gene-set genetic algorithm with the escape operation is proposed. The modified gene-set genetic algorithm can consider both the escape from local optima and the search for global optima. Experiments on three problems are also made to show the effectiveness of the modified genetic algorithm.

A dynamic-edge ACS algorithm for continuous variables problems

AbstractAnt colony systems (ACS) have been successfully applied to solving optimization problems. Especially, they are efficient and effective in finding nearly optimal solutions to discrete search spaces. When the solution spaces of the problems to be solved are continuous, it is not so appropriate to use the original ACS to solve it. This paper thus proposes a dynamic-edge ACS algorithm for solving continuous variables problems. It can dynamically generate edges between two nodes and give a pheromone measures for them in a continuous solution space through distribution functions. In addition, it maps the encoding representation and the operators of the original ACS into continuous spaces easily. The proposed algorithm thus provides a simple and standard approach for applying ACS to a problem that has a continuous solution space, and retains the original process and characteristics of the traditional ACS. Several constrained functions are also evaluated to demonstrate the performance of the proposed algorithm.