Numeral is a crucial part of financial documents. In order to understand the detail of opinions in financial documents, we should not only analyze the text, but also need to assay the numeric information in depth. Because of the informal writing style, analyzing social media data is more challenging than analyzing news and official documents. FinNum is a task for fine-grained numeral understanding in financial social media data - to identify the linking between the target cashtag and the target numeral.
In order to understand the numeral information in-depth, we proposed a task in NTCIR-14 to disambiguate the meaning of the numerals in financial social media data. However, only understanding the meanings of numerals is not enough for practical uses, because there may have more than one cashtag in a financial tweet. Understanding their semantic roles in the financial social media data is needed when mining fine-grained opinions toward a certain target.
Along this line, we design another novel task for fine-grained numeral understanding in financial social media data, called numeral attachment, which aims to detect the attached target (i.e., cashtag) of the numeral. That is, we attempt to understand that the numeral is attached to which cashtag in a tweet. For example, there are two cashtags and one numeral in (T1). The numeral "36.50" is related to $BEXP, instead of $KOG.
(T1) $KOG Took a small position- hopefully a better outcome than getting kneecapped by $BEXP selling itself dirt cheap at 36.50