A Rule Based Feature Selection Approach for Target Classification in Wireless Sensor Networks with Sensitive Data Applications

Abstract

One of the important issues faced in the domain of target classification in wireless sensor networks is the restricted lifetime of individual sensors, caused by limited battery capacity. Although the base station usually has sufficient energy supply and computational power, it is often deemed to be the object of enemy invading hostile terrain. Hence, minimizing energy consumption of sensors while maintaining a given classification accuracy is a key problem in this research area, especially for sensitive data applications. This paper proposes a rule based feature selection approach rather than all-features approach that aims at increasing the energy efficiency of the system without losing much classification accuracy. In experiments, the feasibility and effectiveness of our approach are demonstrated empirically.

Development of Rule-Based Feature Extraction in Multi-label Text Classification

“…The feature extraction and feature selection is the most often used methods in the study of text classification for reducing the number of high-dimensional data [11]. There are many types of feature selections that can be used to handle high dimensions, such as rule-based feature selection to improve efficiency in built systems without reducing the value of accuracy that should be obtained [12].…”

Model-Based Sensitivity Analysis on Aerosol Optical Thickness Prediction

“…However, few researches were focused on the validation of input attributes, or, say, feature selection, for regression models. Although feature selection has been studied in classification tasks [14][15][16][17], few works have been implemented in regress tasks. To some degree, too many input attributes could be redundant or noisy for accurate model prediction.…”