Improving generalization of MLPs with multi-objective optimization
نویسندگان
چکیده
This paper presents a new learning scheme for improving generalization of multilayer perceptrons. The algorithm uses a multi-objective optimization approach to balance between the error of the training data and the norm of network weight vectors to avoid over"tting. The results are compared with support vector machines and standard backpropagation. ( 2000 Elsevier Science B.V. All rights reserved.
منابع مشابه
Decisior Implementation in Neural Model Selection by Multi-objective Optimization
This work presents a new learning scheme for improving generalization of Multilayer Perceptrons (MLPs). The proposed Multi-objective algorithm (MOBJ) approach minimizes both the sum of squared error and the norm of network weight vectors to obtain the Pareto-optimal solutions [1]. Preliminar results are shown in [3]. Since the Pareto-optimal solutions are not unique, we need a decision phase in...
متن کاملAn Approach to Reducing Overfitting in FCM with Evolutionary Optimization
Fuzzy clustering methods are conveniently employed in constructing a fuzzy model of a system, but they need to tune some parameters. In this research, FCM is chosen for fuzzy clustering. Parameters such as the number of clusters and the value of fuzzifier significantly influence the extent of generalization of the fuzzy model. These two parameters require tuning to reduce the overfitting in the...
متن کاملThe Quasi-Normal Direction (QND) Method: An Efficient Method for Finding the Pareto Frontier in Multi-Objective Optimization Problems
In managerial and economic applications, there appear problems in which the goal is to simultaneously optimize several criteria functions (CFs). However, since the CFs are in conflict with each other in such cases, there is not a feasible point available at which all CFs could be optimized simultaneously. Thus, in such cases, a set of points, referred to as 'non-dominate' points (NDPs), will be...
متن کاملMulti Objective Optimization Placement of DG Problem for Different Load Levels on Distribution Systems with Purpose Reduction Loss, Cost and Improving Voltage Profile Based on DAPSO Algorithm
Along with economic growth of countries which leads to their increased energy requirements,the problem of power quality and reliability of the networks have been more considered andin recent decades, we witnessed a noticeable growing trend of distributed generation sources(DG) in distribution networks. Occurrence of DG in distribution systems, in addition tochanging the utilization of these sys...
متن کاملUrban Land-Use Allocation By A Cell-based Multi-Objective Optimization Algorithm
Allocating urban land-uses to land-units with regard to different criteria and constraints is considered as a spatial multi-objective problem. Generating various urban land-use layouts with respect to defined objectives for urban land-use allocation can support urban planners in confirming appropriate layouts. Hence, in this research, a multi-objective optimization algorithm based on grid is pr...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
- Neurocomputing
دوره 35 شماره
صفحات -
تاریخ انتشار 2000