10th EAI International Conference on Bio-inspired Information and Communications Technologies (formerly BIONETICS)

Research Article

Fuzzy Logic Training for Predicting Age of Rats

Download
  • @INPROCEEDINGS{10.4108/eai.22-3-2017.152398,
        author={Douglas Eric Dow and Isao Hayashi},
        title={Fuzzy Logic Training for Predicting Age of Rats},
        proceedings={10th EAI International Conference on Bio-inspired Information and Communications Technologies (formerly BIONETICS)},
        publisher={ACM},
        proceedings_a={BICT},
        year={2017},
        month={3},
        keywords={machine learning fuzzy logic supervised training tuning python age aging rat muscle maximum force muscle function weka},
        doi={10.4108/eai.22-3-2017.152398}
    }
    
  • Douglas Eric Dow
    Isao Hayashi
    Year: 2017
    Fuzzy Logic Training for Predicting Age of Rats
    BICT
    ACM
    DOI: 10.4108/eai.22-3-2017.152398
Douglas Eric Dow1,*, Isao Hayashi2
  • 1: Wentworth Institute of Technology
  • 2: Kansai University
*Contact email: dowd@wit.edu

Abstract

Physiological measurements may contain data with nonlinear relations, non-normal distribution and large signal-to-noise ratio. Fuzzy logic has been utilized to analyze and classify physiological data. Age in mammals is reflected in physiological properties, such as skeletal muscle function. A fuzzy logic algorithm with self-tuning mechanism was developed in this study to make a model from physiological measurements, and predict age. The system was developed using scalar number sets having linear and nonlinear relations. Then the system was applied toward data of body mass and skeletal muscle function to predict the age of rats as a scalar value. The algorithm was developed using the python programming language. The results of the developed fuzzy logic system were compared with other machine learning algorithms using the Weka platform. The developed fuzzy logic model had a lower mean for relative absolute error (RAE) for the tested set of linear and nonlinear relations compared to the results of the tested machine learning algorithms in Weka. For prediction of rat age, the RAE of the fuzzy logic system was 22% compared with values of 23-33% for the other tested algorithms. Further testing and development of the fuzzy logic system on physiological data relations will be necessary to verify these promising results.