All these contracts about investment products have one feature in common: They all contain, somewhere, a number that is the named the “maximum trading amount” or, in other word, the maximum value of one “share”.

The objective of the text mining exercise is to create a solution (with TIMi) that automatically extracts this number: the “maximum trading amount” (as well as some other information such as date of creation, expiration date, country of emission, etc.) from each of the contracts. This is a difficult task because the contracts are written by thousands of different law firms, each time following a completely different layout (that can change over time, and sometimes several times in a day).

This task is still different from automated document classification, which is also something that TIMi can easily do.