A discretization algorithm based on information distance criterion and ant colony optimization algorithm for knowledge extracting on industrial database

Abstract

Discretization algorithms have played an important role in data mining, which is widely applied in industrial control. Since the current discretization methods can not accurately reflect the degree of the class-attribute interdependency of the industrial database, a new discretization algorithm, which is based on information distance criterion and ant colony optimization algorithm(ACO), is proposed. The paper analyses the information measures of the interdependence between two discrete variables, and an improved information distance criterion is generated to evaluate the class-attribute interdependency of the discretization scheme. In the algorithm, The ACO is applied to detect the optimal discretization scheme, and a new pheromone matrix is defined on the construction of the optimization, and an effective heuristic values assignment approach, which is used with the criterion values of discretization scheme, is proposed. We performed the experiments on a real industrial database. Experiment results verify that the proposed algorithm can produce a better discretization results.

A Feature Discretization Method Based on Fuzzy Rough Sets for High-Resolution Remote Sensing Big Data Under Linear Spectral Model

“…First, we compare our method to four discretization algorithms using a heuristic search strategy on running time and search efficiency. These are the classical GA based on the consistency principle of the decision system in the early stage [58], multivariate discretization based on evolutionary cut points selection (EMD) [53], potential particle swarm optimization (PPSO) algorithm [26], and ACO based on information distance criterion [54]. We then compare the optimal set of breakpoints obtained by our method with the discretization results of the current mainstream supervised discretization algorithms, mainly based on the evaluation of the number of intervals and the consistency of the data.…”

“…It is difficult to solve this kind of problem by traditional methods, and global optimization algorithms are more effective than traditional methods because their group search strategies and calculation methods do not depend on gradient information [52]. Global optimization algorithms for discretization mainly include the genetic algorithm (GA) [53], PSO [26], and ant colony optimization (ACO) [54]. Compared to the GA, the PSO has no crossover and mutation, which makes the operation principle simpler.…”

“…The ACO algorithm can obtain good results when the number of intervals is known, but the number of intervals for discretization must be given in advance. The best number of intervals usually cannot be determined, especially with datasets including a large number of samples, where the method fails because it cannot provide the optimal number of intervals [54], [57]. Therefore, given the characteristics of a high-resolution remote sensing image, we use the improved adaptive genetic algorithm to solve the above constraint optimization problem.…”

“…In the experiment, we set the weight coefficient u of FRSGA at 0.1; the weight factors α and β of GA to 1 and 2.5, respectively; the weight factor of EMD to 0.7; the crossover rate p c and mutation rate p m of FRSGA, GA, and EMD to 0.6 and 0.1, respectively; the weight factor µ of PPSO to 0.5; the control variables of the pheromone and heuristic value of ACO to 1 and 2, respectively; and the volatility ρ of the pheromone trace to 0.1. Since the number of classes of a given training samples was five, the uniform class conditional probability vector p e = 1 5 , 1 5 , 1 5 , 1 5 , 1 5 in the ACO algorithm [54]. Because the smaller the fitness function (F itness ∈ [0, 1]) of EMD is, the higher the discretization quality is.…”

Reinforcement Learning-Based Genetic Algorithm in Optimizing Multidimensional Data Discretization Scheme

“…A PSO algorithm was applied to feature selection based on discretization, which can generate more powerful and compact representations in high-dimensional datasets, and thus achieve better classification performance [31]. ACO [32] is used to solve the problem of discretization of continuous features to obtain more concise decision rules and higher prediction accuracy. RS-GA is a mature discretization method, which uses the individual fitness function based on rough sets to evaluate the uncertainty of an information system in a genetic algorithm and searches for the optimal discretization scheme through individual evolution [33].…”