APPLICATION OF A CASE-BASED REASONING APPROACH TO THE PREDICTION OF MINING DISCHARGE

Case-based reasoning (CBR) has been widely used in many real-world applications, such as quality control, customer support, aircraft maintenance, and knowledgebase help systems in software applications. In general, CBR systems predict behavior by comparing some given, unknown case to a library of past cases with the best matching (i.e., most “similar”) retrieved case(s) serving as an approximate solution to the given problem. Our application here is concerned with predicting the hydrologic response of a mine discharge to precipitation. The mine is located in the northern part of Japan. Prediction of mining discharge rate is important in the management of toxic chemicals contained in mine effluent . The cases are used in this approach are developed from a database consisting of eight years of site precipitation and discharge measurements. In the database, these previous discharges are characterized using a set of attributes that includes time between storms, precipitation amount, and time of year. CBR predicts the mine discharge rate for a new storm case by comparing these attributes with those of previous cases and adapting the discharge rates for similar cases. Preliminary results show that this technique can be used to predict discharge values for different types of storms. In addition, the relative weighting applied to the attributes and scoring methods that determine similarity between cases is critically important. CBR has promise as a quick alternative to more involved rainfall-discharge modeling.