A Modified Grey Prediction Method to Early Manufacturing Data Sets

In the age of global competition, product life cycles are shortened and the volatility increases. Industrial demands no longer follows linear trend which could be forecasted relatively easily. As a consequent, enterprise to survival depends on its capability to accommodating changes in manufacturing environment. Due to the rareness of data in the early stages of manufacturing management, traditional prediction techniques hardly obtain ideal results. The urgency of learning and acquirement of the management knowledge is unavoidable for decision makers. Therefore, small-data-set prediction cannot be ignored and should be taken seriously. Nowadays the grey prediction model is one of the important methods for small-data-set prediction. But its application is limited because the way of building model is fixed and could not be modified by the characteristics of data. This research explores the extra information with data analysis through trend and potency tracking method, and uses the generation of trend and potency value of each datum to develop the adaptive grey prediction model as a tool of small-data-set prediction based on the theory of grey system. The example verification displayed that the presented method establishes the suitable model according to the characteristics of data, and also improves the prediction accuracy of small data sets.