Building Material Price Forecasting Based on Multi-method in China

As building material price is highly correlative to project cost, various method are utilized for price prediction. However, their accuracies and ranges of application are still in question. To increase the prediction reliability, this paper proposes a multi-method-based way and use it to predict the main building material price in Wuhan, China. Various methods, including triple exponential smoothing, grey prediction model(GM (1,1)), grey Verhulst model, polynomial fitting method, are used separately to obtain the optimum one with minimum mean square error and its prediction result is adopted as the final prediction value. The results show that: For the common C10 commercial concrete, the relative error of optimal predicted value is 0.4%; For the common hot rolled round steels, although the overall optimal method is GM (1,1), the polynomial fitting method is most accuracy at some local time points. Therefore, the results fully demonstrate the effectiveness and rationality of multi-method.