Prediction of the yield of biohydrogen under scanty data conditions based on GM(1,N)

• GM(1,N) is used for predicting the yield of biohydrogen.
• The method has high predictability under the condition of scanty data.
• The predicted values are certain in each running of the computer codes.
• The affecting extent of each factor on the yield of biohydrogen can be identified.

Biohydrogen technology is regarded as one of the most promising ways for hydrogen production with the considerations of economic priority and environmental performance. In this study, grey model is used to predict the yield of biohydrogen under scanty data condition. An illustrative case has been studied by the proposed method, and pH, glucose and iron sulfate concentration are used as the independent variables, the yield of biohydrogen is used as dependent variable in the grey prediction model, and 9 groups of data are used as the training samples and another 2 groups of data are used as the test samples, the results show that the proposed method is feasible to predict the yield of biohydrogen under scanty data condition and the effect of the influencing factors on the yield could also be identified. According to the comparison with the results predicted by artificial neural network, it could be concluded that grey model has better predictability with scanty data. This method could be popularized to other biohydrogen systems.