Time Series Forecasting for Big Data - Cyber Academic Group
There are various methods to analyze data. ARIMA is represented by three parameters: degree of autoregressive, degree of integration and degree of moving average. This method is popular due to its statistical properties and the availability of classical models (e.g. Box-Jenkins). The assumption of linearity imposes a major limitation on their use. In recent years other models have been developed such as the use of ANNs and hybrid techniques. In certain situations, such models have shown improvements in prediction over ARIMA models. The objectives of this proposal are to analyze a given “big data set” and develop models to describe relationships and behavior between various data sets.
Mechanical & Aerospace Engineering
Department of Defense Space