1st International Conference on Industrial Networks and Intelligent Systems

Research Article

Work-In-Progress: An intelligent diagnosis influenza system based on adaptive neuro-fuzzy inference system

Download710 downloads
  • @INPROCEEDINGS{10.4108/icst.iniscom.2015.259009,
        author={Chun-Ling Lin and Sheng-Ta Hsieh},
        title={Work-In-Progress: An intelligent diagnosis influenza system based on adaptive neuro-fuzzy inference system},
        proceedings={1st International Conference on Industrial Networks and Intelligent Systems},
        publisher={ICST},
        proceedings_a={INISCOM},
        year={2015},
        month={4},
        keywords={adaptive neuro-fuzzy inference system (anfis) membership function (mf) greedy forward feature selection intelligent diagnosis system},
        doi={10.4108/icst.iniscom.2015.259009}
    }
    
  • Chun-Ling Lin
    Sheng-Ta Hsieh
    Year: 2015
    Work-In-Progress: An intelligent diagnosis influenza system based on adaptive neuro-fuzzy inference system
    INISCOM
    ICST
    DOI: 10.4108/icst.iniscom.2015.259009
Chun-Ling Lin1,*, Sheng-Ta Hsieh2
  • 1: Ming Chi University of Technology
  • 2: Oriental Institute of Technology
*Contact email: ginnylin@mail.mcut.edu.tw

Abstract

This study combines adaptive neuro-fuzzy inference system (ANFIS) with greedy forward feature selection to develop the intelligent diagnosis system. Two different membership functions (MFs), Trapezoidal and Gaussian, are adopted during the training process of ANFIS in order to compare the diagnosis accuracy of Trapezoidal MF with one of Gaussian MF. The comparison of ANFIS values with simulated data indices that adoption of both Trapezoidal and Gaussian MF in proposed system achieve satisfactory accuracy (>96%). Furthermore, the accuracy of ANFIS with Gaussian MF is above 98%. Hence, the intelligent diagnosis system can provide a preliminary result to physicians so that the doctor could quickly and accurately decide whether patient have cold or influenza