Water quality prediction based on probability-combination

The existing combination method for water quality prediction lacks of a framework approach and the probability prediction is always based on an assumption that water quality index follows certain probability distribution.A new method based on probability-combination for water quality prediction was proposed.The method combined the prediction results of different single methods through the use of Odds-Matrix method and it could improve the performances of prediction effectively.It was worth noting that the combination-forecast approach could be extended to new methods.The probability of prediction was established through statistical analysis of historical prediction data and hence the validation of the method was achieved along with interval estimation under certain confidence level.The Probability-Combination method based on gray model group and exponent smoothing was proposed for Jiukengkou,Zhejiang.Experimental results indicated that the combination-forecast approach performs better than single prediction method.The validity of probability establishment could be checked effectively.According to the results,the envelopes of water-quality curve under certain confidence level were derived.The prediction accuracy and stability could be improved effectively and probability results which could be the basis of the decision-making could also be given successfully by probability-combination.Furthermore,as a framework approach,it could add new prediction methods continuously.