Abstract: With the economic development, research of the burden of revenue on the same industry is becoming more and more important.The traditional methods neglected many factors that will affect the result of estimating the burden of revenue for the same industry, so the burden of revenue is often below the normal level.This paper introduced C4.5 Algorithm of Data Mining to avoid the disadvantages of the traditional methods.Further, it improved the C4.5 Algorithm by combining with the knowledge of revenue.At last, this method is proved effective by some instances.

Key words: industrywide tax burden distribution, entropy, data mining