Please use this identifier to cite or link to this item:
https://ptsldigital.ukm.my/jspui/handle/123456789/499855Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor.advisor | Zaidi Isa, Prof. Dr. | - |
| dc.contributor.author | Ahmed R.M. Alsayed (P74943) | - |
| dc.date.accessioned | 2023-10-13T09:35:28Z | - |
| dc.date.available | 2023-10-13T09:35:28Z | - |
| dc.date.issued | 2017-01-19 | - |
| dc.identifier.other | ukmvital:85331 | - |
| dc.identifier.uri | https://ptsldigital.ukm.my/jspui/handle/123456789/499855 | - |
| dc.description | The existence of outliers in panel data may violate the assumption of classical OLS estimator, which might lead to misestimating and inaccurate representation on the relationship between the variables. Nevertheless most of the previous studies have applied OLS estimation regardless the existence of outliers. However, outliers are actual values and it is not recommended to exclude them in the analysis as they may contain significant information. To tackle that problem, different robust estimators have been developed. In addition, it is common in environmental phenomena that the data has heterogeneous problem, which masks the nexus between the variables. On the other hand, the setup of the environmental model between energy consumption (EC), economic growth (GDP) and carbon dioxide emissions (CO2) has received much attention due to the global environmental issue and greenhouse effect. Although different methods have been applied to explain that relationship, but the results were inconsistent which may due to the applied statistical methods. The main objectives of this study include (1) to detect the outliers, leverages and influence points by applying different diagnostic methods (studentized residual, leverage, DfBetas, DfFits, Welsch, Covariance Ratio, Cook's distance and Robust Mahalanobis distance) on the environmental panel data; (2) to suggest the best robust estimator for the environmental model by applying different robust estimators (M, Median, S and MM) against OLS estimator in the presence of outliers in the panel data, then evaluate the models by using out-of-sample forecasting approach; (3) to apply the quantile regression in the environmental model for different quantiles (0.05, 0.25, 0.50, 0.75, and 0.95) against OLS estimator model to overcome the heterogeneous data problem; (4) to compare the results using different transformation forms (natural logarithm and inverse form) in reducing the heteroscedasticity in panel data to the environmental model. (5) to test the environmental Kuznets curve (EKC) hypothesis between CO2 emissions and GDP in developed and developing countries. (6) to investigate the different effect of developed countries versus developing countries on environmental degradation. The panel data consists of 29 countries from two different economic levels, 17 developed versus 12 developing countries. The data takes the annually period from year 1960 to 2008. The results show that the robust Mahalanobis distance method outperforms the other methods in identifying the types of outliers. Moreover, the M-estimator is the best robust estimator could represent the environmental model in the presence of outliers. Furthermore the results from quantile regression indicate that there are different significant impacts on lower quantiles (25th and 50th) from CO2 emission than that on the higher quantiles (75th and 95th) and the OLS model. In addition, the estimated model using inverse form of data could represent the environmental relationship better than using data in natural logarithm form. In conclusion, M-estimator is a robust estimator in handling data with high efficiency and high breakdown point with the existence of different types of outliers. Furthermore, Quantile regression could provide different results at different points of the distribution. On other hand, the GDP and EC contribute to higher environmental degradation particularly in CO2 emissions, thus it is recommended to policy makers to consider the effect of these factors to environment quality in future planning or related to policy decision making.,Certification of Master's/Doctoral Thesis" is not available | - |
| dc.language.iso | eng | - |
| dc.publisher | UKM, Bangi | - |
| dc.relation | Faculty of Science and Technology / Fakulti Sains dan Teknologi | - |
| dc.rights | UKM | - |
| dc.subject | Robust estimation | - |
| dc.subject | Outlier detection | - |
| dc.subject | Panel data | - |
| dc.subject | Environmental science | - |
| dc.subject | Dissertations, Academic -- Malaysia | - |
| dc.title | Robust estimation and outlier detection on panel data with application in environmental science 2017 | - |
| dc.type | Theses | - |
| dc.format.pages | 156 | - |
| dc.identifier.barcode | 002578(2017) | - |
| Appears in Collections: | Faculty of Science and Technology / Fakulti Sains dan Teknologi | |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| ukmvital_85331+SOURCE1+SOURCE1.0.PDF Restricted Access | 579.07 kB | Adobe PDF |  View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.