Ensemble based novel class identification for Class Imbalance under sampled Data
Journal Title: International Journal for Research in Applied Science and Engineering Technology (IJRASET) - Year 2016, Vol 4, Issue 11
Abstract
The Classification of data is represented as research challenge in handling the class imbalance problem. Classification problems are represented by highly unbalanced data sets, in which, the number of samples from one class is much smaller than from another. This is known as class imbalance problem and is often reported as an obstacle for constructing a model that can successfully discriminate the minority samples from the majority samples. Under sampling is a popular method in dealing with class-imbalance problems, which uses only a subset of the majority class and thus is very efficient. The main deficiency is that many minority class examples are ignored. Ensemble based under sampling method is proposed for the class imbalance problem. The class imbalance problem is defined in terms of which the ratio of the majority and minority class cardinalities is inversed. The main idea is to severely under sample the majority class thus creating a large number of distinct training sets using normalized information gain. For each training set we then find a decision boundary which separates the minority class from the majority class using the classifier c5.0. By combining the multiple designs through fusion, we construct a composite boundary between the majority class and the minority class using entropy calculation. Experimental results show that both proposed method class-imbalance learning method out performs state of arts approaches higher in terms of precision, recall and f-measure for disproportionate class sample size for different boundaries.
Authors and Affiliations
Sulfyth. M, Mrs. D. Priyadarshini
Keywords
Related Articles
- EP ID EP22779
- DOI -
- Views 262
- Downloads 3
Augmented Reality for Data Booth
Nowadays people widely use internet for purchasing a home, car, furniture etc. In order to obtain information about that product user prefer pamphlets or leaflets or obtain the information by means of Salesperson. Thoug...
A Survey About MANET
Our aim in this paper to take a survey about MANET. That is by collecting the information of transmission capacity and packet delivery delay in mobile ad hoc networks. In order to achieve the fundamental understanding o...
An Interactive Online Employee Training and Tracking System
Employee training is essential for an organization’s success. When any company recruit students from the campus, it never place directly on working before providing any training to them. The company trains their employe...
Comparison between Cascaded Multilevel inverter and reduced switch multilevel inverter
This paper proposes a comparison between cascaded seven level inverter and nine level inverter using Phase opposition and disposition Pulse Width Modulation control scheme. The number of switches used in the cascaded se...
Implementation of Centralized Logging Feature Using Web Services
Data logging is the technique that involves gathering of log data from one or more applications which produces persistent results at runtime with less effort. This paper describes the server side storage of log data eve...