Ensemble MLP Classifier Design

Abstract

Abstract. Multi-layer perceptrons (MLP) make powerful classifiers that may provide superior performance compared with other classifiers, but are often criticized for the number of free parameters. Most commonly, parameters are set with the help of either a validation set or crossvalidation techniques, but there is no guarantee that a pseudo-test set is representative. Further difficulties with MLPs include long training times and local minima. In this chapter, an ensemble of MLP classifiers is proposed to solve these problems. Parameter selection for optimal performance is performed using measures that correlate well with generalisation error.

Hybrid Correlation and Causal Feature Selection for Ensemble Classifiers

“…Final output will be selected from majority vote of all classifiers of each bootstrap replicate. Bootstrap is the most well-known strategy for injecting randomness to improve generalization performance in multiple classifier systems and provides out-of-bootstrap estimate for selecting classifier parameters [24]. Randomness is desirable since it increases diversity among the base classifiers, which is known to be a necessary condition for improved performance.…”

“…Randomness is desirable since it increases diversity among the base classifiers, which is known to be a necessary condition for improved performance. However, there is an inevitable trade-off between accuracy and diversity known as the accuracy/diversity dilemma [24]. Nevertheless, in causal discovery, there are some disadvantages for BNs learning using Bagging.…”

“…An ensemble classifier or multiple classifier system (MCS) is another wellknown technique to improve system accuracy [24]. Ensembles combine multiple base classifiers to learn a target function.…”

Change detection in remotely sensed images using an ensemble of multilayer perceptrons

“…To handle this problem, instead of using a single MLP, a combination of different MLPs can be used. Research work using ensemble of MLPs [16], [17] is found in the literature.…”

Correlation-Based and Causal Feature Selection Analysis for Ensemble Classifiers

“…An ensemble classifier or multiple classifier system (MCS) is another wellknown technique to improve system accuracy [6]. Ensemble combines multiple base classifiers to learn a target function and gathers their prediction together.…”