Constructing Average Decadal Unemployment Rates on Vector Space Model

There are no official unemployment rates in the United States for the decades before 1940. This work measures the titles of economics papers from 1900 to 2012 on the basis of the retrieved data from the Web of Knowledge. The titles are chronologically grouped into 10 datasets and we perform regression models based on vector space model and the US official 1940s – 2012 unemployment rates by the Bureau of Labor Statistics. We tentatively construct the average unemployment rates for the 4 decades before 1940 and compare them with those estimates by other researchers.