For NTCIR-9 we will continue to use the collections used in NTCIR-8 (see below for a full description of these) with the addition of a Korean collection and another English collection (for the same time period as the Korean collection. At present the collections available are
Japanese Dataset: Mainichi 1998-2001, 2002-2005
English Dataset: Mainichi Daily 1998-2001 Korea Times 1998-2001 (New York Times 2002-2005 and Xinhua English 1998-2001 will be distributed by the LDC.)
Korean Dataset: Hankookilbo 1998-2001 Chosunilbo 1998-2001
Collections used in NTCIR-8 are described here.