| Abstract: | Limiting intrusive customs examinations is recommended under the revised Kyoto Convention. It is also a proposal discussed in the context of World Trade Organization (WTO) trade facilitation negotiations. To limit these intrusive examinations, the more modern governments now intervene at all stages of the customs chain, using electronic data exchange and risk analysis, and focusing their resources on a posteriori inspection.
The Ethiopian Revenues and Customs Authority (ERCA) is one of the pioneers in implementing risk management in its customs processing. There is huge amount of data is being stored and processed daily activity for risk management. Just like other developing countries’ customs office , it does not properly utilize its vast data records in a way that enables it to extract pattern and regularities important for forecasting problems related to risk in advance and take appropriate action. The Ethiopian revenue and customs authority is currently using Statistical calculation but mainly manual risk management scheme for selectivity for risk analysis as part of a process of putting analytics at the core of its business processes.
The problem is to be able to handle this huge amount of data and information in such a way that they can identify what is important and be able to extract it from the accumulated data. It is too complex and voluminous to be processed and analyzed by traditional methods. Now a day, data mining technology is being used as a tool that provides the techniques to transform these mounds of data into useful information which in turn enables to derive knowledge for decision making. A number of data mining techniques and tools are available to perform this task. The researcher considered selective techniques and tools which were used to explore the prevalence of Custom risk channel assignment and develop classification and prediction models.
Thus, the purpose of this study is to investigate the potential applicability of data mining techniques in exploring the prevalence of custom risk management using the data collected from Ethiopian Revenue and custom authority risk management database.
Three machine learning algorithms from WEKA software such as J48 Decision trees (DT), Naïve Bayes (NB) and K nearest neighbor classifiers are adopted to classify custom risk channel records on the basis of the values of attributes “Risk Level”. Initially, a total dataset of 18814
vi | P a g e
records with 13 attributes were collected for the study. In this study CRISP-DM model was used as framework.
Results of the experiments have shown that K nearest neighbor (KNN) classifier has better classification and accuracy performance as compared to Naïve Bayes (NB) and Decision Tree classifier. The model selected in evaluation performance of these classifiers has an accuracy of 92.71 %.Overall, this study has proved that data mining techniques are valuable to support and scale up the efficacy of custom services provision process. |
