大数据时代的经济学研究:数据驱动范式

Economics Research in Big Data Era: Data-driven Paradigm

主流经济学研究一般采用模型驱动研究范式,包括选定模型的组成变量、建立模型的基本假设、模型的模拟与求解、实证检验以及分析结论五个步骤。模型驱动研究范式推动了近代以来经济学的发展,与思辨驱动研究范式一道成为经济学发展史上重要的里程碑。但自21世纪以来,数据出现了爆炸式增长,数据量级已达到了ZB(270B)级别,且还在呈指数级加速增长。数据大爆炸给经济学研究带来了全新的挑战,模型驱动研究范式已难以适应经济学研究的需要。随着计算机技术的迅猛发展,数据驱动研究范式因其可最大限度利用巨量数据所提供的有价值的信息而表现出模型驱动研究范式不可比拟的优势,因而未来的经济学研究必将由数据驱动研究范式所主导。对此,经济研究工作者应做出相应改变,以适应大数据时代的来临。

Traditional economics generally uses the model-driven paradigm, which includes five steps： select model variables, establish assumptions, solve the model, test empirically, and analyze conclusion. Model-driven paradigm promotes the development of economics in modern times, and will become important milestones in the history of economics with speculative-driven research paradigm. After entering the 21st century, there has been an explosive growth in data, and the data magnitude has reached ZB （270B） level, and generated by accelerating exponentially. Big bang to the economic research data brings new challenges, so model-driven research paradigm has been difficult to adapt to the needs of economic research, which requires economic researchers to make appropriate changes to accommodate the arrival of the era of big data. With the rapid development of computer technology, data-driven paradigm can maximize the valuable intormation provided by big data. In the future, data-driven paradigm will domain economic research.