The primary objective of the study is to explore the predictability of herding patterns of foreign institutional investors in the Indian market using high frequency data over a period from January 2003 to June 2014. Herding of an individual stock was measured estimating a simple volume based ratio and persistence of trends was detected using the runs test (Wald and Wolfowitz [1940]) on that ratio. Predictability of herding behavior has been successfully modeled by applying 7 data mining models using various measures of performance. Market regulators may consider our findings to regulate the foreign institutional investors trading to make the financial system more transparent and robust.