China
Africa
Explore chapters and articles related to this topic
En-Fuzzy-ClaF
Published in Neeraj Mohan, Surbhi Gupta, Chuan-Ming Liu, Society 5.0 and the Future of Emerging Computational Technologies, 2022
Sourabh Shastri, Sachin Kumar, Kuljeet Singh, Vibhakar Mansotra
Fuzzy classification is the method of grouping elements into a set of fuzzy whose membership function can be described by the truth values of a fuzzy propositional function. Its simplest form is rule-based fuzzy classifier (or if-then fuzzy system). Let’s take an example of two classes whose fuzzy classifier can be constructed by identifying the if-then rule: IF fever is normal AND cough is normal THEN class is non-COVID-19.IF fever is moderate AND cough is severe THEN class is COVID-19.IF fever is severe AND cough is moderate THEN class is COVID-19.
Classification of Type-2 Diabetes using Bat-based Fuzzy Rule Miner
Published in Ramalingaswamy Cheruku, Damodar Reddy Edla, Venkatanareshbabu Kuppili, Soft Computing Techniques for Type-2 Diabetes Data Classification, 2020
Ramalingaswamy Cheruku, Damodar Reddy Edla, Venkatanareshbabu Kuppili
Fuzzy classification rules are more interpretable and cope better with pervasive uncertainty and vagueness with respect to crisp rules. Because of this fact, fuzzy classification rules are extensively used in classification and decision support systems for disease diagnosis. But, most of the rule mining techniques failed to generate accurate and comprehensive fuzzy rules. This chapter presents a novel approach to extract fuzzy classification rules from bio-inspired-based bat algorithm called RST-BatMiner. The proposed RST-BatMiner integrates the Rough Set Theory (RST) for feature selection and bat algorithm for fuzzy rule extraction [38]. It also introduces a new operator into the encoding scheme of bat to generate comprehensive rules. Moreover, a boosting mechanism is combined with the bat algorithm to increase the accuracy of generated fuzzy rules. The proposed RST-BatMiner is experimentally tested on the PID dataset and results are compared with other bio-inspired-based fuzzy rule miners using accuracy, sensitivity, specificity, number of rules, mean antecedent rule length and mean ruleset size.
Tuberculosis Detection from Conventional Sputum Smear Microscopic Images Using Machine Learning Techniques
Published in Siddhartha Bhattacharyya, Václav Snášel, Indrajit Pan, Debashis De, Hybrid Computational Intelligence, 2019
Rani Oomman Panicker, Biju Soman, M.K. Sabu
(d) Fuzzy classification based method A fuzzy classification based method was proposed by Ghosh et al. [15] in 2016. The method in [15] has a pre-processing stage, feature extraction stage (shape, color and granularity) and fuzzy classification stage; followed by a final classification stage. They proposed a gradient based region growing technique for accurately detecting the contour of TB bacilli. However, their shape-based method in [15] failed in detecting the overlapping or touching TB bacilli. But they claimed that color- and granularity-based methods detected touching bacilli, so they used a majority voting for the final decision making.
Applying a new systematic fuzzy FMEA technique for risk management in light steel frame systems
Published in Journal of Asian Architecture and Building Engineering, 2022
Ali Yeganeh, Moein Younesi Heravi, Seyed Behnam Razavian, Kourosh Behzadian, Hashem Shariatmadar
Fuzzy theory is a computing method using “degrees of truth“ rather than the traditional ”true or false” (1 or 0) Boolean logic that underpins modern computers (Meng Tay and Peng Lim 2006). The concept of fuzzy sets was introduced by Zadeh in 1960s for the first time (Jong, Tay, and Lim 2013). In this approach, a fuzzy set described the concepts of a fuzzy number by using a degree of membership of its elements in a universe of discourse (Sang et al. 2018). Fuzzy numbers defined in the interval [0,1] provide semantics for terms in a linguistic term set, which are represented by MF that can be classified by types of functions. Fuzzy set theory is also used in a fuzzy inference system (FIS) to generate a model between inputs (features in the case of fuzzy classification) and targets (classes in the case of fuzzy classification). Due to the use of FIS, such transition may need a set of fuzzy rules in which gathering a complete one is difficult (Jee, Tay, and Lim 2015; Kerk et al. 2021). Previous researches have indicated all of the above concepts could adopt to the risk analysis due to the capability of fuzzy concept for modelling of uncertainty.
Distributed Probabilistic Fuzzy Rule Mining for Clinical Decision Making
Published in Fuzzy Information and Engineering, 2021
Samane Sharif, Mohammad-R. Akbarzadeh-T
A good agent is defined as an agent with a confidence level above a pre-determined threshold (9). The LPFRBs of good agents are sent to a central part. The union of these LPFRBs, denoted as PFRB and defined by (10), can be used for decision making and classification problems. The proposed classifier employs probabilistic fuzzy rules, while most of the previous fuzzy classification methods, either centralized or distributed, have been designed based on conventional fuzzy rules.
A classification-based fuzzy-rules proxy model to assist in the full model selection problem in high volume datasets
Published in Journal of Experimental & Theoretical Artificial Intelligence, 2022
Angel Díaz-Pacheco, Carlos Alberto Reyes-Garcia
The strategies studied showed great interest in improving the accuracy of proxy models. In the FMS problem, a proxy model is used as a mechanism to decide if a candidate model is promising enough to test it with the true model or if it should be discarded. Through the search’s progression, changes in the meta-dataset also change the importance of features and the suitability of learning algorithms. The FMS paradigm provides an excellent alternative for built flexible proxy models in this ever-changing search space. A good foundation of promising models will allow the bio-inspired mechanisms to converge to optimal zones in the search space. We thought outside the box and, discrimination between potential and disposable models can be addressed as a classification task. An extension of this idea is that a fuzzy-classification algorithm can provide a membership degree of the classified object to considered classes. This membership degree can be used as an upper bound to perform finer discrimination. In this way, the time of the process is invested only in highly potential models, which saves time and leads to better solutions. To get a better understanding of our approach and its difference with other strategies in the literature, Figure 1 (A) and (B) are provided. As can be seen in Figure 1 (A), we can understand the FMS process as a box where a dataset is fed and as output, we get a highly accurate model. However, this process needs to build and test many models until to find the best. With this in mind, we proposed a considerable reduction of this time-consuming process and a way to guide the procedure through the search space. One difference with other proxy-based approaches is that construction of the proxy model is performed with the FMS approach, as observed in Figure 1 (B). Bearing this in mind, we can argue that the major contribution of this work is proposing a way to improve the quality of proxy models and the shift of the paradigm on this field. We transformed this problem from the prediction of the fitness of a provisional model to a discrimination on its potential, ‘to test, or not to test, that is the question’.