Data Mining Technique Analysis

Data Mining Technique Analysis

Thesis Statement

The term data mining is defined as the process of gathering practical and useful information from a set of large data by using the techniques of data analysis in order to help management to make better decisions.

Genetic Algorithm

Genetic algorithm (GA) is an optimization method based on the concepts of natural selection and genetics. In this approach, the variables that characterize the solution presented in the form of a gene in a chromosome. GA operates on a finite set of solutions (population) - generates new solutions as different combinations of parts of the population of solutions using operators such as selection, recombination (crossover) and mutation (Bies, 2006, pp.96-105.). New solutions are positioned in the population according to their position on the surface of the function.

GA suggested the idea of the nature and work of Darwin. It is assumed that if you take two very good solutions in any way to get them out of the new solution, there will be a high probability that the new solution will turn out good or even better. Then we can arrange some environment - population, inhabit its decisions - by individuals, and give them a fight (de Ville, 2006, pp.36-45.). To do this, define a function for which will be determined by the power of individuals - the quality of its proposed solution. Based on this parameter, this can be determined for each individual number left by her descendants, or the likelihood that the individual is left to posterity. Also do not rule out the possibility, when the specimen is too low value of this parameter will die.

Advantages of Genetic Algorithms

They do not require any information about the surface of the answer;

The gaps that exist on the surface of the answer have little effect on the overall efficiency of optimization;

They are resistant to penetration into local optima;

They work well in solving large-scale optimization problems;

Can be used for a wide class of problems;

Simple and transparent in the implementation;

Can be used in problems with a changing environment

Disadvantages of Genetic Algorithms

Not desirable and problematic use of GA:

In the case when you need to find the exact global optimum;

Execution time of the evaluation function is large;

Need to find all solutions of the problem and not one of them;

The configuration is not simple (encoding decision).

Decision Tree Analysis

A decision tree is a diagram that describes the decision-making process by considering the alternatives ...