A data mining-constraint satisfaction optimization problem for cost effective classification

Authors:
Parag C. Pendharkar
Affiliations:
School of Business Administration, Pennsylvania State University at Harrrisburg, Middletown, PA
Venue:
Computers and Operations Research
Year:
2006

Citing 10
Cited 1

Managerial applications of neural networks: the case of bank failure predictions

Management Science
A comparative study of classification algorithms: statistical, machine learning and neural network

Machine intelligence 13
Feature minimization within decision trees

Computational Optimization and Applications
The zero/one multiple knapsack problem and genetic algorithms

SAC '94 Proceedings of the 1994 ACM symposium on Applied computing
Artificial Neural Networks

Artificial Neural Networks
Sequential Decision Models for Expert System Optimization

IEEE Transactions on Knowledge and Data Engineering
Dynamic Programming and Strong Bounds for the 0-1 Knapsack Problem

Management Science
Mathematical Programming for Data Mining: Formulations and Challenges

INFORMS Journal on Computing
An empirical study of impact of crossover operators on the performance of non-binary genetic algorithm based neural approaches for classification

Computers and Operations Research
Cost-sensitive classification: empirical evaluation of a hybrid genetic decision tree induction algorithm

Journal of Artificial Intelligence Research

A GRASP method for building classification trees

Expert Systems with Applications: An International Journal

Quantified Score

Hi-index	0.01

Visualization

Abstract

We propose a data mining-constraint satisfaction optimization problem (DM-CSOP) where it is desired to maximize the number of correct classifications at a lowest possible information acquisition cost. We show that the problem can be formulated as a set of several binary variable knapsack optimization problems, which are solved sequentially. We propose a heuristic hybrid simulated annealing and gradient-descent artificial neural network (ANN) procedure to solve the DM-CSOP. Using a real-world heart disease data set, we show that the proposed hybrid procedure provides a low-cost and high-quality solution when compared to a traditional ANN classification approach.The massive proliferation of very large databases in organizations makes it necessary to design cost effective and efficient data mining systems. This paper proposes a data mining constraint satisfaction optimization problem, which provides a high quality cost effective solution for a binary classification problem.