Multicollinearity Effect in Regression Analysis: A Feed Forward Artificial Neural Network Approach

Obite, Chukwudi Paul; Olewuezi, N. P.; Ugwuanyim, Geoffrey U.; Bartholomew, Desmond Chekwube

doi:10.9734/ajpas/2020/v6i130151

Cited by 49 publications

(21 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“… is the bias that can be interpreted as the intercept in a linear regression, A final transformation 0  is applied to the output (Obite et al, 2020), and the quadratic error function given in equation ( 5) was used in this study to determine the weights.…”

Section: Methodsmentioning

confidence: 99%

Modeling US Dollar and Nigerian Naira Exchange Rates During COVID-19 Pandemic Period: Identification of a High-performance Model for New Applications

Obite

Nwosu

Bartholomew

2021

jmss

View full text Add to dashboard Cite

This study modeled the US Dollar and Nigerian Naira exchange rates during COVID-19 pandemic period using a classical statistical method – Autoregressive Integrated Moving Average (ARIMA) – and two machine learning methods – Artificial Neural Network (ANN) and Random Forest (RF). The data were divided into two sets namely: the training set and the test set. The training set was used to obtain the parameters of the model, and the performance of the estimated model was validated on the test set that served as new data. Though the ARIMA and random forest performed slightly better than the neural network in the training set, their performance in the test set was poor. The neural network with 5 nodes in the input layer, 5 nodes in the hidden layer and 1 node in the output layer (ANN (5,5,1)) performed better on the new data set (test set) and is chosen as the best model to forecast for future USD to NGN exchange rate. The information from the high-performance model (ANN (5, 5, 1)) for modeling the USD to NGN exchange rate will assist econometric trading of the currencies and offer both speculative and precautionary assistance to individuals, households, firms and nations who use the currencies locally and for international trade.

show abstract

Section: Methodsmentioning

confidence: 99%

Modeling US Dollar and Nigerian Naira Exchange Rates During COVID-19 Pandemic Period: Identification of a High-performance Model for New Applications

Obite

Nwosu

Bartholomew

2021

jmss

View full text Add to dashboard Cite

show abstract

“…In order to achieve this, Akpanta and Iwueze's [23] method of using Bartlett's transformation [24] was used. The regression line obtained was = 1.07602 − 3.99102 (See Appendix II for full result of the regression analysis).…”

Section: Results From Sarimamentioning

confidence: 99%

“…To ensure that the seasonal artificial neural network would converge quickly [22] the input and target variables were scaled. The input and target variables were scaled using the Min-Max normalization approach thus:…”

Section: (D) Forecastingmentioning

confidence: 99%

See 1 more Smart Citation

Comparison of the Performance of the SANN, SARIMA and ARIMA Models for Forecasting Quarterly GDP of Nigeria

Nwokike

Okereke

2021

ARJOM

View full text Add to dashboard Cite

This research aimed at modelling and forecasting the quarterly GDP of Nigeria using the Seasonal Artificial Neural Network (SANN), SARIMA and Box-Jenkins models as well as comparing their predictive performance. The three models mentioned earlier were successfully fitted to the data set. Tentative architecture for the SANN was suggested by varying the number of neurons in the hidden layer while that of the input and output layer remained constant at 4. It was observed that the best architecture was when the hidden layer had 10 neurons and thus SANN (4-10-4) was chosen as the best. In fitting the ARIMA/SARIMA models, the Augmented Dickey Fuller (ADF) test was used to check for stationarity. Variance stabilization and Stationarity were achieved after logarithm transformation and first regular differencing. The ARIMA/SARIMA model with lowest AIC, BIC and HQIC values was chosen as the best amongst the competing models and fitted to the data. The adequacy of the fitted models was confirmed observing the correlogram of the residuals and the Ljung-Box Chi-Squared test result. The SANN model performed better than the SARIMA and ARIMA models as it had a Mean Squared Error value of 0.004 while SARIMA and ARIMA had mean squared errors of 0.527 and 0.705 respectively. It was concluded that the SANN which is a non-linear model be used in modelling the quarterly GDP of Nigeria. Hybrid models which combine the strength of individual models are recommended for further research.

show abstract

“…Test for multi-collinearity. Multi-collinearity refers to the problem when independent variables in a multiple regression model are highly correlated with one another (Obite et al, 2020;Wonsuk et al, 2014;Zhang et al, 2020). The presence of multi-collinearity can lead to wider confidence intervals and less reliable probability values (P values) for the parameter estimates of independent variables, drastically reducing statistical significance and predictive power of the regression model (Paul, 2006;Zhou and Huang, 2018).…”

Section: Logistic Regression Modelmentioning

confidence: 99%

The influencing factors of Harbin (China) residents’ satisfaction with municipal solid waste treatment

et al. 2020

View full text Add to dashboard Cite

China is experiencing an enormous increase in municipal household solid waste (MHSW) generation and is facing multiple problems associated with the treatment of MHSW. This paper analyses factors affecting residents’ satisfaction with MHSW treatment performance. Six factors were identified by the Delphi method: (a) pick-up frequency by waste collection vehicles, (b) fund supply situation, (c) charging standard for waste treatment, (d) waste bin arrangement, (e) laws and regulations, (f) publicity and education. We examine the significance of these six influencing factors, estimating binary logistic regression models. Data for this study are derived from the survey responses of 469 households in Harbin, one of the largest cities in northeast China. The results indicate that ‘pick-up frequency by waste collection vehicles’ is ranked the first and most important determinant of Harbin residents’ satisfaction with MHSW treatment; this is closely followed by ‘publicity and education’. The third and fourth significant influencing factors, respectively, are ‘fund supply situation’ and ‘charging standard for waste treatment’. The last two factors are ‘laws and regulations’ and ‘waste bin arrangement’. By understanding the influence of various factors on residents’ satisfaction, this study aims to help in designing an effective waste management system to reduce the cost of MHSW management, and to raise the residents’ satisfaction with municipal solid waste treatment. Based on the research findings, we advocate that establishing a reasonable waste transport (pick-up) system as well as strengthening publicity and education of waste management are key to improving residents’ satisfaction with the MHSW treatment performance.

show abstract

Multicollinearity Effect in Regression Analysis: A Feed Forward Artificial Neural Network Approach

Cited by 49 publications

References 3 publications

Modeling US Dollar and Nigerian Naira Exchange Rates During COVID-19 Pandemic Period: Identification of a High-performance Model for New Applications

Modeling US Dollar and Nigerian Naira Exchange Rates During COVID-19 Pandemic Period: Identification of a High-performance Model for New Applications

Comparison of the Performance of the SANN, SARIMA and ARIMA Models for Forecasting Quarterly GDP of Nigeria

The influencing factors of Harbin (China) residents’ satisfaction with municipal solid waste treatment

Contact Info

Product

Resources

About