Hi Divya,
Thermocouples have a non-linear transfer function but for a limited temperature you can assume a linear transfer function. I would like to refer you to an old app note AN225. In Appendix A page 8 you will see a method for calculating a linear Seebeck coefficient for a limited temperature range. You will need to know the temperature range of your cold junction. The more limited this range is the less error you will see. Once you calculate this coefficient you can use it to determine the resistance value for R2. I would probably keep R1 at 100k. The figure we show is for a type J thermocouple and for the cold junction temperature compensation the slope used was 50.5uV/C from taking 505/100k*10mV/C. You should consult the manufacturer of your cable to get the thermoelectric voltage table as manufacturing impurities can affect the slope slightly. Also make sure that the output of the thermocouple is going to a high impedance low bias current amplifier as the bias current can cause additional error (especially if it has a temperature coefficient usually doubles every 10C or so for some cases).
Take care