Ren-Jieh Kuo1, Man-Hsin Huang2, Wei-Che Cheng3, Chih-Chieh Lin4, Yung-Hung Wu5. 1. Department of Industrial Management, National Taiwan University of Science and Technology, No. 43, Section 4, Kee-Lung Road, Taipei 106, Taiwan. Electronic address: rjkuo@mail.ntust.edu.tw. 2. Hon Hai Precision Industry Co., Ltd., No. 2, Zihyou St., Tucheng Dist., New Taipei City 236, Taiwan. 3. Department of Industrial Management, National Taiwan University of Science and Technology, No. 43, Section 4, Kee-Lung Road, Taipei 106, Taiwan. 4. Department of Urology, Taipei Veterans General Hospital, Institute of Clinical Medicine, 201, Section 2, Shih-Pai Road, Taipei 112, Taiwan; Department of Urology, School of Medicine, National Yang-Ming University, Taipei, Taiwan. 5. Superintendent Office, Taipei Veterans General Hospital, 201, Section 2, Shih-Pai Road, Taipei 112, Taiwan.
Abstract
OBJECTIVE: This study intends to develop a two-stage fuzzy neural network (FNN) for prognoses of prostate cancer. METHODS: Due to the difficulty of making prognoses of prostate cancer, this study proposes a two-stage FNN for prediction. The initial membership function parameters of FNN are determined by cluster analysis. Then, an integration of the optimization version of an artificial immune network (Opt-aiNET) and a particle swarm optimization (PSO) algorithm is developed to investigate the relationship between the inputs and outputs. RESULTS: The evaluation results for three benchmark functions show that the proposed two-stage FNN has better performance than the other algorithms. In addition, model evaluation results indicate that the proposed algorithm really can predict prognoses of prostate cancer more accurately. CONCLUSIONS: The proposed two-stage FNN is able to learn the relationship between the clinical features and the prognosis of prostate cancer. Once the clinical data are known, the prognosis of prostate cancer patient can be predicted. Furthermore, unlike artificial neural networks, it is much easier to interpret the training results of the proposed network since they are in the form of fuzzy IF-THEN rules. These rules are very important for medical doctors. This can dramatically assist medical doctors to make decisions.
OBJECTIVE: This study intends to develop a two-stage fuzzy neural network (FNN) for prognoses of prostate cancer. METHODS: Due to the difficulty of making prognoses of prostate cancer, this study proposes a two-stage FNN for prediction. The initial membership function parameters of FNN are determined by cluster analysis. Then, an integration of the optimization version of an artificial immune network (Opt-aiNET) and a particle swarm optimization (PSO) algorithm is developed to investigate the relationship between the inputs and outputs. RESULTS: The evaluation results for three benchmark functions show that the proposed two-stage FNN has better performance than the other algorithms. In addition, model evaluation results indicate that the proposed algorithm really can predict prognoses of prostate cancer more accurately. CONCLUSIONS: The proposed two-stage FNN is able to learn the relationship between the clinical features and the prognosis of prostate cancer. Once the clinical data are known, the prognosis of prostate cancerpatient can be predicted. Furthermore, unlike artificial neural networks, it is much easier to interpret the training results of the proposed network since they are in the form of fuzzy IF-THEN rules. These rules are very important for medical doctors. This can dramatically assist medical doctors to make decisions.
Keywords:
Fuzzy neural network; Optimization version of an artificial immune network; Particle swarm optimization algorithm; Prognosis; Prostate cancer