Classification of Imbalanced Data Using a Modified Fuzzy-Neighbor Weighted Approach
Journal Title: International Journal of Intelligent Engineering and Systems - Year 2017, Vol 10, Issue 1
Abstract
Classification of imbalanced datasets is one of the widely explored challenges of the decade. The imbalance occurs in many real world datasets due to uneven distribution of data into classes, i.e. one class has more instances while others have a few that results in the biased performances of traditional classifiers towards the majority class with large number of instances and ignorance of other classes with less data. Many solutions have been proposed to deal with this issue in various crisp and fuzzy methods. This paper proposes a new hybrid fuzzy weighted nearest neighbor approach to find better overall classification performance for both minority and majority classes of imbalanced data. Benefits of neighbor weighted K nearest neighbor approach i.e. assignment of large weights to small classes and small weights to large classes are merged with fuzzy logic. Fuzzy classification helps in classifying objects more adequately as it determines that how much an object belongs to a class. Experimental results exhibit the improvements in classification of imbalanced data of different imbalance ratios in comparison with other methods.
Authors and Affiliations
Harshita Patel
Keywords
Related Articles
- EP ID EP229394
- DOI 10.22266/ijies2017.0228.07
- Views 125
- Downloads 0
A Novel QoS Trust Computation in MANETs Using Fuzzy Petri Nets
Evaluation of node’s trust value is certainly advantage in mobile Adhoc networks (MANETs) where the applications run efficiently by involving trustable nodes only. The proposed method “A Novel QoS Trust computation in MA...
Opposition Learning-Based Grey Wolf Optimizer Algorithm for Parallel Machine Scheduling in Cloud Environment
Cloud computing is a novel developing computing paradigm where implementations, information, and IT services are given over the internet. The parallel-machine scheduling (Task-Resource) is the important role in cloud com...
A Secure Client Aware Certification for Mobile Cloud Offloading Decision
The advancements in smartphones with excellent feasibility and networking capabilities paved the way for rising leap in mobile communication, but their limitations incline users to next level paradigm called Mobile Cloud...
An Effective Implementation of Dual Path Fused Floating-Point Add-Subtract Unit for Reconfigurable Architectures
Reconfigurable architectures have provided a low cost, fast turnaround platform for the development and deployment of designs in communication and signal processing applications. The floating point operations are used in...
Ant Colony and Tolerable Delay Based Modified Virtual Backoff Algorithm for Wireless Sensor Networks
The MAC layer plays a major role in the enhancement of the performance of any network. Hence this paper proposes a modified virtual backoff algorithm which queues the packets based on the delay that application can toler...