A Novel Algorithm for Scaling up the Accuracy of Decision Trees
Journal Title: International Journal on Computer Science and Engineering - Year 2010, Vol 2, Issue 2
Abstract
Classification is one of the most efficient and widely used data mining technique. In classification, Decision trees can handle high dimensional data, and their representation is intuitive and generally easy to assimilate by humans. The area under the receiver operating characteristic curve, AUC is one of the recently used measures for calculating the performance of a classifier.In this paper, we presented two novel decision tree algorithms namely C4.45 and C4.55, aimed to improve the AUC value over the C4.5, which is a state-of-the-art decision tree algorithm. The empirical experiments conducted on 42 benchmark datasets have strongly indicated that C4.45 and C4.55 has significantly outperformed C4.5 on the AUC value.
Authors and Affiliations
Ali Mirza Mahmood , K. Mrutunjaya Rao , Kiran Kumar Reddi
Optimization of Composite Plates Based on Imperialist Competitive Algorithm
Imperialist Competitive Algorithm (ICA) is a new optimization algorithm that inspired by socio-political process of imperialistic competition. In this paper an optimization methodology for the design of composite plates...
Optimized design of BCD adder and Carry skip BCD adder using reversible logic gates
Reversible logic is very essential for the construction of low power, low loss computational structures which are very essential for the construction of arithmetic circuits used in quantum computation, nano technology an...
INCREASING THE EFFICIENCY OF THE SOFTWARE RCHITECTURE RECOVERY THROUGH SPANNING TREE BASED MAXIMAL GRAPH MINING TECHNIQUE
This paper represents a technique for recovering the Software Architecture based on Graph Pattern Matching by the help of mining techniques. enerally Software Architecture is represented in terms of graphs with set of...
NODE ADDRESSING SCHEMES FOR SCALABLE ROUTING IN HIERARCHICAL WSN
Wireless Sensor Network has a capacity of communication computation with node in a network. The sensor performs many significant tasks in a selective area. In this paper proposing the new idea to improve issues on node a...
On high order methods for solution of nonlinear equation
The objective of this paper is two folds, first to derive a new series of third order methods for solving non-linear equation f(x)=0 involving third, fourth and fifth derivatives of f and secondly to compare the existing...