Improving the Performance of Artificial Neural Networks via Instance Selection and Feature Dimensionality Reduction

Abstract

Abstract-This paper presents a hybrid approach with two phases for improving the performance of training artificial neural networks (ANNs) by selection of the most important instances for training, and then reduction the dimensionality of features. The ANNs which are applied in this paper for validation, are included Multi-Layer Perceptron (MLP) and Neuro-Fuzzy Network (NFN). In the first phase, the Modified Fast Condensed Nearest Neighbor (MFCNN) algorithm is used to construct the subset with instances very close to the decision boundary. It leads to achieve the instances more useful for training the network. And in the second phase, an Ant-based approach to the supervised reduction of feature dimensionality is introduced, aims to reduce the complexity, and improve the accuracy of learning the ANN. The main purpose of this method is to enhance the classification performance by improving the quality of the training set. Experimental results illustrated the efficiency of the proposed approach.Index Terms-Artificial neural networks, instance selection, feature selection, ant colony optimization, EDA, MFCNN. I. INTRODUCTIONThe learning procedure has different steps including, constructing a Training Set (TS), training the network, testing the behavior of the network on the new instances. The first step consists of editing and condensing the data. The most important objective of any abstraction technique is to obtain a TS with the best performance in classification process. To build an optimal TS, we have to select the instances and features in an appropriate way to guarantee attaining to the best result.The whole data is located in a matrix which its rows show the instances and the columns are the features. The instance selection schemes generally try to reduce the number of rows in entire dataset with no loss in the classification performance [1]. Non-parametric K-Nearest Neighbor (KNN) rule [2], predicts the label of a new instance by computing a similarity scale between the selected one and all other instances in TS. Condensed Nearest Neighbor rule (CNN) [3], was introduced for the nearest neighbor classification method and tried to build a condensed subset from the whole set of examples. The basic idea of this algorithm is that if an example is not correctly classified, this may occur because it is a boundary instance, and it should be considered in the selected condense set.Fast set S that is initialized with centroid of each class in TS. Then, until the TS is not completely classified with condensed set S, the algorithm selects for insertion a representative of the misclassified points of each Voronoi (Voronoi of point P∈S is the set of all instances in TS that are closer to P than to any other instance in S) cell induced by the current subset. Edited Nearest Neighbor (ENN) removes noisy instances, as well as close border points, leaving smoother decision boundaries [5]. It edits out those prototypes that misclassified using KNN rule. Allknn [6] is based on the ENN method. This algorithm, for k=1 to m, marks...

Automatic Simplified Symbolic Analysis of Analog Circuits Using Modified Nodal Analysis and Genetic Algorithm

“…An exhaustive search strategy can generate the best solution. [13][14][15][16] However, evaluation of all possible subsets is impractical even for small-size circuits, due to computational complexity. 16 It has been shown that swarm intelligence algorithms (e.g., GA, particle swarm optimization, ant colony optimization, etc.)…”

“…can e±ciently be applied for NP-hard problems. [13][14][15][16] Therefore, GA is used for the simpli¯cation of analog circuits in this paper in order to achieve the best performance.…”

“…In this approach, the mutation probability P m is considered to be decreased linearly from P max m to P min m during execution of algorithm, as seen in Eq. (13). The larger the P m , the more symbolic terms will be complemented via mutation operator.…”

“…The idea of applying a swarm intelligence algorithm for the symbolic simpli¯cation of analog circuits in this paper was inspired by our previous works on three popular binary selection problems [13][14][15] : Instance Selection problem, 13 Feature Selection problem, 14 and Knapsack problem. 15 Although these problems were not used in the case of analog circuits, they are structurally similar to the symbolic simpli¯cation problem.…”

On the use of local and global search paradigms for computer-aided diagnosis of breast cancer

“…Using all the aforementioned attributes may not result in the most accurate CAD system because there may be some irrelevant or noisy attributes that hinder the system's ability to distinguish between healthy and cancer patients correctly. Therefore, it is crucial to adopt an appropriate attribute selection strategy in order to select the most important subset of attributes that can be later used to classify patients accurately (Abroudi et al, 2013). Numerous CAD systems for early detection of breast cancer have been presented and they have used different data as well as different metrics to measure the system classification performance (Table 1).…”