Improved Adaptive Neuro-Fuzzy Inference System for HIV/AIDS Time Series Prediction

Improving accuracy in time series prediction has always been a challenging task for researchers. Prediction of time series data in healthcare such as HIV/AIDS data assumes importance in healthcare management. Statistical techniques such as moving average (MA), weighted moving average (WMA) and autoregressive integrated moving average (ARIMA) models have limitations in handling the non-linear relationships among the data. Artificial intelligence (AI) techniques such as neural networks are considered to be better for prediction of non-linear data. In general, for complex healthcare data, it may be difficult to obtain high prediction accuracy rates using the statistical or AI models individually. To solve this problem, a hybrid model such as adaptive neuro-fuzzy inference system (ANFIS) is required. In this paper, we propose an improved ANFIS model to predict HIV/AIDS data. Using two statistical indicators, namely, Root Mean Square Error (RMSE) and Mean Absolute Error (MAE), the prediction accuracy of the proposed model is compared with the accuracies obtained with MA, WMA, ARIMA and Neural Network models based on HIV/AIDS data. The results indicate that the proposed model yields improvements as high as 87.84% compared to the other models.