ABSTRACT
Efficiency and precision in prediction of
Chlorophyll-a using this model is still a pandemic among researchers, due to
the natural conditions in ocean water systems itself, which involved chemical,
biological and physical processes and interaction among them may affect the
model performance drastically. Thus, to overcome this problem as well as to
improve the strength of MLR, we proposed a hybrid approach, i.e., an Artificial
Neural Network to the MLR coins as Artificial Neural Network-Multiple Linear
Regression (ANN-MLR). To investigate the performance of the proposed model, we
compared Multiple Linear Regression (MLR), Artificial Neural Network (ANN) and
proposed hybrid Artificial Neural Network and Multiple Linear Regression
(ANN-MLR) in the prediction of chlorophyll-a (chl-a) concentration by
statistical measurement which are MSE and MAE. Achieving our objectives of
study, we used 4 parameters, i.e. temperature (°C), pH, salinity (ppt), DO
(ppm) at the Offshore Kuala Terengganu, Terengganu, Malaysia. The results
showed that our proposed model can improve the performance of the model as
compared to ANN and MLR due to small errors generated, error reduced, and
increased the correlation coefficient for all parameters in both MSE and MAE,
respectively. Thus, this result indicated that our proposed model is efficient,
precise and almost perfect correlation as compared to ANN and MLR.
Share and Cite:
Lola, M. , Ramlee, M. , Gunalan, G. , Zainuddin, N. , Zakariya, R. , Idris, M. and Khalil, I. (2016) Improved the Prediction of Multiple Linear Regression Model Performance Using the Hybrid Approach: A Case Study of Chlorophyll-a at the Offshore Kuala Terengganu, Terengganu.
Open Journal of Statistics,
6, 789-804. doi:
10.4236/ojs.2016.65065.