Combining spatial and semantic label analysis

Abstract

Labels in spreadsheets can be exploited for finding errors in spreadsheet formulas. Previous approaches have either used the positional information of labels or their interpretation as dimension for checking the consistency of formulas.In this paper we demonstrate how these two approaches can be combined. We have formalized a combined reasoning system and have implemented a corresponding prototype system. We have evaluated the system on the EUSES spreadsheet corpus. The evaluation has demonstrated that adding a syntactic, spatial analysis to a dimension inference can significantly improve the rate of detected errors.

Reasoning about spreadsheets with labels and dimensions

“…Previous research [8] has shown that the majority of dimensions are found on only one axis. Specifically, only 128 out of the 567 tables in the investigated spreadsheets contained dimensions on more than one axis.…”

“…While the method described in the previous section has been shown to be effective at detecting errors in spreadsheets [8], it is but one way to combine these labels and dimensions. In fact, this system was designed only as an add on to the dimension inference system, and so in some cases it ignores the search for label errors to focus instead on identifying dimension errors.…”

“…In previous work [8] we have seen that some label errors were missed because some of the labels were exclusively used for dimension analysis. As that system was designed to be simply an add on to the dimension inference tool, this was not an immediate concern.…”

“…In recent work [8] we have designed an integrated system that combines label-based reasoning with dimension inference. This approach was achieved by gathering both label and dimension information about a cell.…”

“…In this paper we will describe several different options, along with the original integrated label and dimension analysis [8], that can be pursued to check a spreadsheet for label and dimension errors. We will evaluate these different systems by running them on the 487 dimension and formula contained sheets in the EUSES corpus, as determined in , [4] and determine the efficiency and correctness for each combination.…”