A Machine Learning Approach for Layout Inference in Spreadsheets

Abstract

Abstract:Spreadsheet applications are one of the most used tools for content generation and presentation in industry and the Web. In spite of this success, there does not exist a comprehensive approach to automatically extract and reuse the richness of data maintained in this format. The biggest obstacle is the lack of awareness about the structure of the data in spreadsheets, which otherwise could provide the means to automatically understand and extract knowledge from these files. In this paper, we propose a classification approach to discover the layout of tables in spreadsheets. Therefore, we focus on the cell level, considering a wide range of features not covered before by related work. We evaluated the performance of our classifiers on a large dataset covering three different corpora from various domains. Finally, our work includes a novel technique for detecting and repairing incorrectly classified cells in a post-processing step. The experimental results show that our approach delivers very high accuracy bringing us a crucial step closer towards automatic table extraction.

Table Identification and Reconstruction in Spreadsheets

“…In this paper, we look at worksheets whose cells were previously classied by our method [10]. We assign a label to each non-empty cell.…”

“…Additionally, we give priority to Headers over Attributes. It is fair to claim that Headers represent more secure fences, since less misclassication involve this label compared to Attributes [10]. Another details is that of S4 and S5 being executed only after all the types of fences are processed by steps S1-S3.…”

“…This translates into 465 individual worksheets. The annotations were performed by experts from our group, using a tool we developed in our previous work [10]. Each non-empty cell was assigned one of the ve predened labels (see Figure 1) Attribute regions (as mentioned in Section 3, these regions form the base of tables).…”

“…Koci, E., Thiele, M., Romero, O., Lehner, W. Table identification [10] the understanding of data maintained in these les.…”

“…In a previous paper [10], we have proposed a machine learning approach for layout inference. We focused on the level of individual cells, considering a large number of features not covered by related work.…”