DATA MINING TECHNIQUE FOR STRUCTURAL STRENGTH MONITERING SYSTEM METHODOLOGIES

摘要

Natural Hazards are the current issues to effect the Building infra-Structures. But the difficult task is to know the status of Building infra-structures, in terms of life time, strength, quality and status of damages in the structures of the buildings. As on today, any automatic methodology are not available for predict or estimate the damages in the infra-structures of the buildings. So essential maintenance can be required. Structural Strength Monitoring System (SSMS) are one of the automatic Monitoring System for satisfying the requirements like predicating damages, classification of damage structures. When SVM classifier is used for calculating the strength of the structures, it cannot effectively handle large data received from big structures but its accuracy for handling small structures is good and ANN classifier handles large amount of data gathered from sensors but its accuracy is low. To design a framework which can effectively handle large amount of data and to improve accuracy rate a hybrid algorithm combining the features of Rough set Support vector machine (RS-SVM) classified structures and Rough set Artificial Neural Network (RS-ANN) is proposed. Structural Strength Monitoring System (SSMS) utilizing Wireless Sensor Systems (WSS) has picked up research interest because of its capacity to reduce the expenses related with the establishment and upkeep of SSMS frameworks. This methodology contains, the combination of both feature subset reduction methods like Rough set theory (RST), Mutual Information (MI) etc. and Classification methods in Data-Mining like SVM, ANN etc.