Study on Loss Function: Postgraduate Study Opportunity (URGENTLY NEEDED ASSISTANT - monthly allowance provided)

I'm currently doing a research on loss function to be used in system identification. I invite anyone who satisfies the requirements of the university (Admission Requirements (utem.edu.my)) to apply position as research assistant and also become postgraduate student (MSc). Provided below is the synopsis:

When it comes to applying an information criterion in system identification process, many evaluations of the optimality of a model are based on the number of variables. Among the well-known ones are Akaike information criterion and Bayesian information criterion. Even though they account for both variance contribution and bias contribution, they were developed for a general class of problems. Furthermore, they require information of number of samples and number of parameters which are inadequate in evaluating and making optimal selection of model structure. It is noted that magnitude of parameter may play a big role in choosing whether a term is significant enough to be included in a model. An information criterion that has recently been proven to address the matter effectively is called parameter magnitude-based information criterion 2 (PMIC2) that had been tested on linear and nonlinear autoregressive with exogenous variable (ARX and NARX) model (Samad and Nasir, 2017, 2018a). This study intends to apply the new information criterion on a more complex model that is able to provide better predictive accuracy. A notable advance may be achieved by testing it on linear and Nonlinear AutoRegressive Moving Average with eXogenous variable (ARMAX and NARMAX) model. This research will simulate the effectiveness of the criterion (and improve it where applicable) towards better model structure selection. This shall be tested using computational software on a number of simulated systems of various lag orders and number of term/variables and, a real plant data. Furthermore, these findings shall be compared with the calculated value using other well-known loss functions. The output is an application of such information criterion in selecting an optimum model structure for a more accurate model representation. With a more accurate model representation, better control strategy may be applied through various related industries towards achieving better production operation.

Anyone interested in taking this further together (as research assistant, co-researcher or acquaintance for data testing), contact me. The application can be done through this address (Home Page -Student Online Portal (utem.edu.my)). The research requires high interest in mathematics and fast learning of computer programming preferably MATLAB.