Application of a noisy data classification technique to determine the occurrence of flashover in compartment fires

Design of a two-stage fuzzy classification model

“…Abonyi et al (2003) gives a 95.46% CR with the decision-tree method. Lee et al (2006) figure out that the CR is 95% on a hybrid Artificial Neural Network model. Our proposed fuzzy classification model can achieve a 99.33% CR with K = 7, and with K = 6 the CR is 98.67%.…”

“…The design of a classification system can be accomplished through a knowledge-based approach or a data-driven approach. Due to the importance of classification problems, many different methods, such as fuzzy logic (Subasi, 2006), neural network (Lee, Lee, Lim, & Tang, 2006;Lin, Yeh, Liang, & Chung, 2006;Mads, Pedersen, Hansen, & Larsen, 1996), GAs (Guo, Li, & Kuo, 2002;Setnes & Roubos, 2000), and a statistical approach (Chen & Hsu, 2006), have been developed to design classification systems.…”

An incremental adaptive neural network model for online noisy data regression and its application to compartment fire studies

“…The superior performance of these techniques compared with those from traditional models has also been addressed. However, the application of ANN techniques to determine the consequences of building fires is still very limited [48][49][50], and this area of work constitutes the main study of this paper.…”

Intelligent-based Structural Damage Detection Model

“…The corrupted output can be separated into clean and noisy components as shown in Eq. (9) where the variable ε is the symmetrically distributed noise with zero mean:…”

“…This compression scheme also facilitates the removal of symmetrically distributed noise embedded in the training samples. The GRNNFA model was successfully applied in fire safety engineering [9,10]. In this article, the damage detection method follows the approach of multi-dimensional classification.…”