Self-Fuzzification Method according to Typicality Correlation for Classification on tiny Data Sets

Abstract

-This article presents a self-fuzzification method to enhance the settings of a Fuzzy Reasoning Classification adapted to the automated inspection of wooden boards. The supervised classification is made thanks to fuzzy linguistic rules generated from small training data sets. This study especially answers to a double industrial need about the pattern recognition in wooden boards. Firstly, few samples are available to generate the recognition model. This aspect makes lesser efficient compilation methods like neural networks in terms of recognition rates. Secondly, the settings of the classification method must be simplified, because the users are not experts in fuzzy logic. In this article, two points are presented. The first part demonstrates the generalization capability of the presented classification method in comparison to more classical algorithms. In the second part, we propose a new automatic method of parameter fuzzification, by using the typicality correlation coefficients of each class.

Fuzzy Rule Iterative Feature Selection (FRIFS) with Respect to the Choquet Integral Apply to Fabric Defect Recognition

“…Then, from the ratio Corr/Xcorr, which characterizes the interclasses similarity, the number of terms is determined. Their positions are obtained by calculating the mean value of the samples belonging to the considered output classes [10]. The main interest takes place in the automatic adaptation of the fuzzification step which makes the tuning of the system easier.…”

“…: Fuzzy Reasoning Classifier) [12] is based on fuzzy linguistic rule mechanism; which is well adapted to our industrial application. Indeed, it presents a very good and efficient generalisation from a few sample set and is able to provide gradual membership for output classes [10]. Its satisfactory behaviour has been shown in [10] by several comparisons with other classifiers such as k Nearest Neighbor (k-NN), Neural Networks (NN) or Support Vector machine (SVM).…”

“…Indeed, it presents a very good and efficient generalisation from a few sample set and is able to provide gradual membership for output classes [10]. Its satisfactory behaviour has been shown in [10] by several comparisons with other classifiers such as k Nearest Neighbor (k-NN), Neural Networks (NN) or Support Vector machine (SVM). This implemented algorithm for the fuzzy recognition method is a supervised learning mechanism which can be decomposed into three parts: Input fuzzification (features of the characteristic vector), Fuzzy rule generation and Rule adjustment (Fig.…”

“…If no information is processed that is the input space is not covered by rule set; the output gives an "unknown defect".The chosen classifier is based on Ishibuchi's algorithm which provides an automatic rule generation step [19]. There are many methods, which automatically provide fuzzy rules according to data set such as a genetic algorithm [20], but the Ishibushi's algorithm is quite simple and gives better results [10]. Moreover, its inference mechanism follows the Larsen model, which is better than the Mamdani model, because the Product is more adapted than the Minimum for the manipulation of several premises [20].…”

Using fuzzy rules for network behavior identification: application for differentiated services in an Ethernet network

“…The major issue with these methods is that they require a large amount of simulation data. Thus, we choose the Fuzzy Rule Classifier (FRC) method, 42 which is well adapted for modeling a system with a reduced dataset. The performance of this approach in terms of generalization has been compared to other classifiers (K-NN, NN, SVM, GA, and DTM) 4 .…”