基于少数据云推理的短时交通流预测模型

针对短时交通流所存在的不确定性即模糊性与随机性特点和准周期规律，提出基于灰色关联分析和少数据云推理的短时交通流预测模型.首先，针对短时交通流的准周期规律，运用灰色关联分析提取不同日期相同时段历史序列中最相似序列；其次，提出少数据逆向云算法，建立交通流序列一维云推理机制；最后综合利用历史云及当前云生成预测云，用于短时交通流实时预测.实例分析表明，预测精度良好，能够有效实现短时交通流的实时预测.该模型解决了少数据条件下正向云参数确定问题，降低了数据处理工作量，开拓了云模型在短时交通流中的应用.

Short-term Traffic Flow Forecasting Model Based on Few Data Cloud Inference

Concerning the fuzziness and randomness characteristics and quasi-periodic regularity in shortterm traffic flow, a short-term traffic flow forecasting model is developed using grey relational analysis and few data cloud inference. Firstly, according to quasi-periodic regularity in short-term traffic flow, the most similar sequence in the history is extracted by gray relational analysis. Then, the backward cloud algorithm of few data is developed, which establishes the mechanism of one-dimensional cloud reasoning of traffic flow sequence. Finally, the prediction cloud is generated by a one-dimensional cloud inference of historical and current information. The results show that this model is used in forecasting short-term traffic flows and the accuracy is considerably improved. This proposed model solves the confirmation of forward cloud parameters under few data conditions, reducing the data processing workload and extending the application scope of the traditional cloud model.