In this paper we investigate automatic data- text alignment, i.e. the task of automatically aligning data records with textual descrip- tions, such that data tokens are aligned with the word strings that describe them. Our meth- ods make use of log likelihood ratios to esti- mate the strength of association between data tokens and text tokens. We investigate data- text alignment at the document level and at the sentence level, reporting results for sev- eral methodological variants as well as base- lines. We find that log likelihood ratios pro- vide a strong basis for predicting data-text alignment.

Click Here to View the Paper In XML Format