Abstract
In recent years, rapid development E-commerce has occurred in China. E-commerce and longtail economy are playing an increasingly important role in the national economy. How to accurately forecast the sales of E-commerce website is valuable. On the basis of building theoretical frame of keywords search queries and online shopping transaction volume, this paper took Taobao.com - the largest E-commerce retail website in China-as an example, composed the keywords into composite index by stepwise synthesis, then analyzed the relationship between the sales of different categories of goods and the corresponding composite index through the regression model, made a prediction test, and finally got the forecast results of daily sales for different categories goods. Compared the forecast results to the actual data, the Mean Absolute Percentage Error is fewer than nine percent.