Advances and Applications in Statistics
Volume 52, Issue 4, Pages 267 - 280
(April 2018) http://dx.doi.org/10.17654/AS052040267 |
|
SHORT AND LONG-TERM PREDICTION BASED ON COMPLEX STATISTICS AND ARTIFICIAL INTELLIGENCE TECHNIQUES
Mona Mustafa El Biely
|
Abstract: This paper compares ANFIS (adaptive network-based fuzzy inference system) and ARIMAX (auto regressive integrated moving average model with exogenous variables), where the goal is to minimize the prediction error for stock price as time series. The statistical method (ARIMAX) uses the immediate past so its accuracy tends to decrease with increasing time. The data are stock prices for Healthcare Services Group, Inc. (HCSG) from January 2012 to December 2016. To determine the characteristics of ANFIS that best suited the target forecasting, several ANFIS models were trained, tested and compared, different types and number of membership functions to generate the initial fuzzy inference system (FIS) were analyzed. This paper examines the performance of ARIMAX and ANFIS models in forecasting stock prices from (13/10/2015) to (13/12/2016). Moreover, the ANFIS is performed better than ARIMAX in long-term prediction. ANFIS and ARIMAX are provided by many computer packages, including MATLAB R2016a and R-3.4.0. |
Keywords and phrases: ANFIS model, ARIMA model, stock prices, long and short prediction. |
|
Number of Downloads: 439 | Number of Views: 3776 |
|