www.sciedu.ca/air Artificial Intelligence Research 2014, Vol. 3, No. 4 ORIGINAL RESEARCH A hybrid knowledge discovery system for oil spillage risks pattern classification Udoinyang Godwin Inyang 1 , Oluwole Charles Akinyokun *2 1 Department of Computer Science, Faculty of Science, University of Uyo, Uyo, Nigeria 2 Department of Physical Sciences, Landmark University, Omu-aran, Nigeria Received: July 30, 2014 Accepted: September 18, 2014 Online Published: November 4, 2014 DOI: 10.5430/air.v3n4p77 URL: http://dx.doi.org/10.5430/air.v3n4p77 Abstract The complexity and the dynamism of oil spillages make it difficult for planners and responders to produce robust plans towards their management. There is need for an understanding of the nature, sources, impact and responses required to prevent or control their occurrence. This paper develops an intelligent hybrid system driven by Sugeno-Type Adaptive Neuro Fuzzy Inference System (ANFIS) for the identification, extraction and classification of oil spillage risk patterns. Dataset consisting of 1008 records was used for training, validation and testing of the system. Result of sensitivity analysis shows that Cause, Location and Type of spilled oil have cumulative significance of 85.1%. Optimal weights of Neural Network (NN) were determined via Genetic Algorithm with hybrid encoding scheme. The Mean Squared Error (MSE) of NN training is 0.2405. NN training, validation and testing results yielded R > 0.839 in all cases indicating a strong linear relationship between each output and target data. Rule pruning was performed with support (15%) and confidence (10%) minimum thresholds and antecedent-size of 3. The performance of the ANFIS was evaluated with eight different types of membership functions (MFs) and two learning algorithms. The model with triangular MF gave the best performance among all other given models while hybrid-learning algorithm performed better than back propagation algorithm. The ANFIS model reported in the paper adopted triangular MF and hybrid learning algorithm for the predication and classification of oil spillage risk patterns. Average training and testing MSE of the model is 0.414315 and 0.221402 respectively. The knowledge mining results show that ANFIS based systems provide satisfactory results in the prediction and classification of oil spillage risk patterns. Key Words: ANFIS, Triangular membership function, Fuzzy logic, Rule interestingness, Oil spillage patterns 1 Introduction Developments in Information and Communication Tech- nology have resulted in huge data repositories for analysis and management by public and private sectors of the world economy. A major requirement for a modern knowledge driven society is the effective and efficient management of data held in these repositories and transforming them into information and knowledge. [1] This gives rise to the need for improved techniques, procedures and tools to aid humans in the automatic and intelligent collection and analysis of huge data sets. Knowledge Discovery (KD) effectively uncovers hidden but subtle patterns from large and diverse datasets and out performs traditional statistical techniques. [2, 3] Data mining, a major stage in the KD process, is the analysis of datasets that are observational, aiming at finding out hid- den relationships among datasets and summarizing the data in such a manner that is both understandable and useful to the users. [4] Some of the intelligent tools for data mining include Neural Networks (NNs), Fuzzy Logic (FL), Ants * Correspondence: Oluwole Charles Akinyokun; Email: akinwole2003@yahoo.co.uk; Address: Department of Physical Sciences, Landmark Uni- versity, Omu-aran, Nigeria Published by Sciedu Press 77