Abstract: The author introduces the boost weights of evidences method (BoostWofE), a new and generalized model of weights of evidence which relaxed the requirement of the ordinary weights of evidence model on conditional independency among evidential layers. It integrates evidences in sequential order and calculates the weights for the first evidence and the conditional weights for subsequent evidences. The BoostWofE is developed to calculate the conditional weights in just the same way used to calculate the ordinary weights but on the basis of a progressively weighted training sample. The method is applied to a case study of involving four layers of binary patterns and 20 training samples of mineral deposits and occurrences for prediction of gold mineral deposits in Nova Scotia, Canada. The results obtained by the ordinary weights of evidence method show strong bias due to significant conditional dependency of evidences with respect to the training sample, but results of posterior probability and subsequently estimation of number of mineral deposits obtained using the new BoostWofE method show significant reduction of bias.

Key words: weights of evidence model, conditional independency, posterior probability, information integration, mineral resources, quantitative assessment, mineral deposit prediction