Elsevier

Applied Soft Computing

Volume 25, December 2014, Pages 242-252
Applied Soft Computing

AGFS: Adaptive Genetic Fuzzy System for medical data classification

https://doi.org/10.1016/j.asoc.2014.09.032Get rights and content

Highlights

  • Adaptive Genetic Fuzzy System (AGFS) for optimizing rules and membership functions.

  • Examined the system with seven datasets using quantitative and qualitative analysis.

  • Achieved average accuracy of 87%.

  • AGFS can be functional to different types of application related to classification.

Abstract

A Genetic Fuzzy System (GFS) is basically a fuzzy system augmented by a learning process based on a genetic algorithm (GA). Fuzzy systems have demonstrated their ability to solve different kinds of problems in various application domains. Currently, there is an increasing interest to augment fuzzy systems with learning and adaptation capabilities. Two of the most successful approaches to hybridize fuzzy systems with learning and adaptation methods have been made in the realm of soft computing. The GA can be merged with Fuzzy system for different purposes like rule selection, membership function optimization, rule generation, co-efficient optimization, for data classification. Here we propose an Adaptive Genetic Fuzzy System (AGFS) for optimizing rules and membership functions for medical data classification process. The primary intension of the research is 1) Generating rules from data as well as for the optimized rules selection, adapting of genetic algorithm is done and to explain the exploration problem in genetic algorithm, introduction of new operator, called systematic addition is done, 2) Proposing a simple technique for scheming of membership function and Discretization, and 3) Designing a fitness function by allowing the frequency of occurrence of the rules in the training data. Finally, to establish the efficiency of the proposed classifier the presentation of the anticipated genetic-fuzzy classifier is evaluated with quantitative, qualitative and comparative analysis. From the outcome, AGFS obtained better accuracy when compared to the existing systems.

Introduction

Computational Intelligence methods such as artificial neural networks, fuzzy logic, and genetic algorithms (GAs) are admired research subjects, since they can deal with composite engineering problems which are complicated to explain by classical methods [1]. In the Computational Intelligence community hybrid approaches have gained considerable attention. One of the most popular approaches is the hybridization among Fuzzy Logic and GA that leads to genetic fuzzy systems (GFSs) [2], [7], [8], [60], [61], [64]. A GFS is fundamentally a fuzzy system augmented, which includes GAs, genetic programming and evolutionary strategies among other evolutionary algorithms (EAs) by a learning process derived from evolutionary computation and GA [5]. The foremost focus here is on the problem of obtaining a compact and precise fuzzy rule-based model from examination data. Commonly, in data-driven fuzzy modeling proposal, the Takagi–Sugeno–Kang (TSK)-type fuzzy model [3] and Mamdani fuzzy model are adopted [3].

A Genetic Algorithm (GAs) is one such method that has established a lot of concentration in recent journalism, owing its recognition to the opportunity of searching irregular and high-dimensional solution spaces. FRBCSs – Fuzzy Rule-Based Classification Systems [3], [4], [14], [15] are productive and recognized tools in the machine learning framework, since they can present an interpretable model for the end user [5], [6], [7], [8]. Including anomaly intrusion detection [9], image processing [10], there are numerous real applications in which FRBCSs have been engaged. In most of these areas, the existing or useful data consist of a high number of patterns (instances or examples) and/or variables. The inductive learning of FRBCSs suffers from exponential growth of the fuzzy rule search space in these circumstances. With the capability to investigate a large search space for appropriate solutions only requiring the performance measure, the routine definition of an FRBS can be seen as an optimization or search problem, and where GAs are a recognized and extensively used global search method. Furthermore, to incorporate a priori knowledge the generic code structure and independent presentation features of GAs make them appropriate candidates. For scheming FRBSs over the last few years these capabilities extended the use of GAs in the expansion of a wide range of approaches.

In recent times, for efficient classification purpose, researchers make the hybridization of fuzzy with other technique like, neural network, genetic algorithm and decision tree. The reason behind combining the fuzzy with other techniques is that the inability of fuzzy classifier in better classification without providing the proper fuzzy rules. Neural network, decision tree and genetic algorithms are used to create the fuzzy rules and the rules generated from those techniques are then specified to the fuzzy rule base for classification process. But, to decrease the computational complexity and the number of algorithms used, rule generation and optimization within a single algorithm is required. Accordingly, the neural network and decision tree is not much proficient of selecting and optimizing the rules. Similarly to produce the optimized rules since it follow the evolutionary computation process, the genetic algorithm can be good choice.

There are numerous techniques of soft computing family among, which Fuzzy Logic (FL) and Genetic Algorithm (GA) are the most significant techniques. By means of fuzzy logic based technique, imprecision, uncertainty and human oriented knowledge representation is possible; but yet self-learning and generalization of rules in not possible. Robust general purpose search algorithms that use principles motivated by natural population genetics to develop solutions to the trouble are Genetic Algorithms (GA). GA produces flexibility to interface with present models and simple to hybridize [14], [15]. With intellectual information systems where genetic fuzzy methodology has been effectively implemented, among which Hybridized techniques of GA and Fuzzy are very functional real world applications. In all domains such as power system, data mining, image processing etc. the Genetic Fuzzy system is applied. The following list explains some of the problems solved by means of Genetic Fuzzy Systems.

  • 1.

    Diagnostic system for disease such as myocardial infraction, breast cancer, diabetes, dental development age prediction, abdominal pain, etc. [41], [42], [43], [44].

  • 2.

    A trading system with GA for optimized fuzzy model [45].

  • 3.

    For optimizing social regulation policies [46].

  • 4.

    Self-integrating knowledge-based brain tumor diagnostics system [47].

  • 5.

    Classification of rules in dermatology data sets for medicine [48].

  • 6.

    Integrating design stages for engineering using GA [49].

  • 7.

    Multilingual question classification through GFS [50].

  • 8.

    University admission process through evolutionary computing [51].

  • 9.

    Genetic mining for topic based on concept distribution [52].

  • 10.

    Intelligent web miner with Neural-Genetic-Fuzzy approach [53].

  • 11.

    Extraction of fuzzy classification rules with genetic expression programming [54].

  • 12.

    Integrated approach for intrusion detection system using GA [55].

In our work, to resolve the disadvantages of fuzzy system we have joint the genetic algorithm with fuzzy system. The basic association of the paper is given as follows: Section 2 represents the review of related works and Section 3 presents the problem description and solution. Section 4 illustrates the contributions made in the work. Section 5 discusses the proposed genetic fuzzy classifier with the applicable diagram and mathematical equations. Results and discussion is specified in section 6 and the practical implication of the proposed genetic-fuzzy classifier is discussed in Section 7. The conclusion is summed up in Section 8.

Section snippets

Review of related works

This section represents the detailed survey of the fuzzy genetic system accessible in the literature particularly, in the standard journals. After cautiously analyzing the literature, by means of the genetic algorithm the drawbacks of the fuzzy classifier are solved. So, every author considered some of the drawbacks of the fuzzy system and the specific drawbacks are solved using the genetic programming method. Our research is frequently concentrated on rule generation, rule selection and

Problem definition and solution

  • 1)

    This section represents the detailed description about the problem statements of Genetic Fuzzy System (GFS). Here we stated 4 different problems and how our proposed methodology will give solutions in detailed manner. The 4 problems statements are, Expert's knowledge is needed to define the rule base.

  • 2)

    Experts knowledge is needed to design the membership function.

  • 3)

    Selection of less and suitable rules.

  • 4)

    Selection of rules with constrained length.

Proposed Adaptive Genetic Fuzzy System for classification

By combining Genetic algorithm (GA) with the fuzzy set, we have offered an adaptive genetic fuzzy system (AGFS). Optimized Rule generation by means of GA, and classification using fuzzy classifier are the process of the proposed AGFS, which is divided into two main steps. To decide on the additional important rules that is then stored to the fuzzy rule base, primarily the training data set is specified to genetic algorithm. In the next step, for classification task the fuzzy system is

Results and discussion

This section represents the experimental outcome and its discussion of the proposed genetic fuzzy classification system. In the subsequent section experimental set up and description about the datasets taken for experimentation is described. To make sure the performance of the proposed system the quantitative, qualitative and comparative analysis has been performed.

Conclusion and future scope

By the combination of Genetic Algorithm with the fuzzy set we have initiated a new classifier namely, Adaptive Genetic Fuzzy classifier. At this point, rule optimization was done by AGA with the aid of new systematic addition and the classification was passed out by fuzzy classifier. Here the frequency of occurrence of the rules in the training data is considered as the fitness for AGA. In conclusion, by means of quantitative, qualitative and comparative analysis the achievement of the proposed

References (66)

  • J. Sanz et al.

    A genetic tuning to improve the performance of Fuzzy Rule-Based Classification Systems with Interval-Valued Fuzzy Sets: Degree of ignorance and lateral position

    Int. J. Approx. Reason.

    (2011)
  • M. Cococcioni et al.

    On reducing computational overhead in multi-objective genetic Takagi–Sugeno fuzzy systems

    Appl. Soft Comput.

    (2011)
  • F.J. Berlanga et al.

    GP-COACH: Genetic Programming-based learning of COmpact and ACcurate fuzzy rule-based classification systems for High-dimensional problems

    Inform. Sci.

    (2010)
  • N.R. Cazarez-Castro et al.

    Fuzzy logic control with genetic membership function parameters optimization for the output regulation of a servomechanism with nonlinear backlash

    Expert Syst. Appl.

    (2010)
  • R. Martínez et al.

    Optimization of interval type-2 fuzzy logic controllers for a perturbed autonomous wheeled mobile robot using genetic algorithms

    Inform. Sci.

    (2009)
  • J. Alcalá - Fdez et al.

    Learning the membership function contexts for mining fuzzy association rules by using genetic algorithms

    Fuzzy Sets Syst.

    (2009)
  • A.G. Evsukoff et al.

    Design of interpretable fuzzy rule-based classifiers using spectral analysis with structure and parameters optimization

    Fuzzy Sets Syst.

    (2009)
  • L. Sánchez et al.

    Genetic learning of fuzzy rules based on low quality data

    Fuzzy Sets Syst.

    (2009)
  • C.F. Juang et al.

    A self-generating fuzzy system with ant and particle swarm cooperative optimization

    Expert Syst. Appl.

    (2009)
  • M.Z. Jahromi et al.

    A proposed method for learning rule weights in fuzzy rule-based classification systems

    Fuzzy Sets Syst.

    (2008)
  • H. Ishibuchi et al.

    Analysis of interpretability-accuracy tradeoff of fuzzy systems by multiobjective fuzzy genetics-based machine learning

    Int. J. Approx. Reason.

    (2007)
  • C.H. Tsang et al.

    Genetic-fuzzy rule mining approach and evaluation of feature selection techniques for anomaly intrusion detection

    Pattern Recog.

    (2007)
  • E. Zhou et al.

    Fuzzy classifier design using genetic algorithms

    Pattern Recog.

    (2007)
  • Y. Maeda et al.

    Fuzzy adaptive search method for parallel genetic algorithm with island combination process

    Int. J. Approx. Reason.

    (2006)
  • C.S. Lee et al.

    A genetic fuzzy agent using ontology model for meeting scheduling system

    Inform. Sci.

    (2006)
  • Y.C. Hu

    Determining membership functions and minimum fuzzy support in finding fuzzy association rules for classification problems

    Knowledge-Based Syst.

    (2006)
  • P. Pulkkinen et al.

    Fuzzy classifier identification using decision tree and multiobjective evolutionary algorithms

    Int. J. Approx. Reason.

    (2008)
  • H. Wang et al.

    Multi-objective hierarchical genetic algorithm for interpretable fuzzy rule-based knowledge extraction

    Fuzzy Sets Syst.

    (2005)
  • A. Konar

    Computational Intelligence: Principles, Techniques and Applications

    (2005)
  • O. Cordón et al.

    Genetic Fuzzy Systems, Evolutionary Tuning and Learning of Fuzzy Knowledge Bases

    (2001)
  • M. Russo

    FuGeNeSys—A fuzzy genetic neural system for fuzzy modeling

    IEEE Trans. Fuzzy Syst.

    (1998)
  • R. Alcalá et al.

    A proposal for the genetic lateral tuning of linguistic fuzzy systems and its interaction with rule selection

    IEEE Trans. Fuzzy Syst.

    (2007)
  • J. Casillas et al.

    Fuzzy-XCS: A Michigan Genetic Fuzzy System

    IEEE Trans. Fuzzy Syst.

    (2007)
  • Cited by (0)

    View full text