Respiratory motion prediction by fuzzy logic approaches: ANFIS vs

Purpose Respiratory motion prediction is a chaotic time series prediction problem. In this study, respiratory motion predictability from 12 traces from breast cancer patients is examined by using Adaptive Neuro-Fuzzy Inference System (ANFIS) and Interval Type-2 Non Singleton Fuzzy System (IT2NSFLS). Methods Free breathing data curves were obtained from Real Time Position Management system (RPM system) from 12 breast cancer patients. The sampling rate was 25 Hz. Two different methods (ANFIS and IT2NSFLS) were tested over a prediction interval of 20 seconds. Results Average root mean square error (RMSE) values were calculated to be 0.1268 and 0.0538 for prediction by ANFIS and IT2NSFLS, respectively. It is also seen that IT2NSFLS is more robust and precise than ANFIS for predicting respiratory motion in breast cancer patients. Purpose In radiation therapy, it is known that the target motion often affects the conformability of the therapeutic dose distribution delivered to thoracic and abdominal tumors. Tumor motions can not only be associated with patient's stochastic movements and systematic drifts, but also involve internal movements caused by such as respiration and cardiac cycles [1]. The target motion often affects the conformability of the therapeutic dose distribution delivered to thoracic and abdominal tumors, and thus tumor motion monitoring systems have been developed. Even we can observe tumor motion accurately; however, radiotherapy systems may inherently have mechanical and computational delays to be compensated for synchronizing dose delivery with the motion. For solving the delay problem, we have compared two different fuzzy logic reasoning approaches: Adaptive Neuro Fuzzy Inference System (ANFIS) and Interval Type-2 Non Singleton Fuzzy Logic System (IT2NSFLS). Methods We describe briefly the two approaches tested for respiratory motion prediction. a) ANFIS: Combines the ability of Neural Network (NN) to learn, with the Fuzzy Logic (FL) to reason in order to form a hybrid intelligent system called ANFIS (Adaptive Neuro Fuzzy Inference System). The goal of ANFIS is to find a model or mapping that will correctly associate the inputs (initial values) with the target (predicted values). Fuzzy inference system (FIS) is a knowledge representation where each fuzzy rule describes a local behaviour of the system. If we view a FIS as a feed-forward network structure where the primary inputs and intermediate results are being sent around to compute the final output, then we can apply the same back-propagation principle for learning as in neural networks. The network structure that implements FIS and employs hybrid-learning rules …