After keyword extraction, input documents and the document index are fetched to the Encryption block.
Validation of input documents is at the heart of the CMLLite approach.
The input documents are ready for encryption once the indexing is finished.
The input documents below should all result in a ValidationResult of VALID and the ValidationReport MUST NOT contain info elements.
There is no such restriction on info elements and these may occur for input documents that otherwise conform completely to the convention.
Rather than rewrite JUMBO [6] and other CML libraries, we designed a set of rules for conformant input documents and tools to process validation.
As previously discussed, OSCAR3 required that input documents be converted into SciXML before processing can occur, using the document formatting as a base against which annotations for identified named entities can be referenced-whether as inline or standoff annotations.
The input documents below should all result in a ValidationResult of VALID and the ValidationReport should contain a single info element and MUST NOT contain either error or warning elements.
The result of an XPath expression is a selection of nodes from the input documents, or an atomic value, or more generally, any sequence allowed by the data model.
Table 1 Overview of unsupervised keyphrase extraction approaches Category Methods Information Sources Co-occurrence based CM [3], SW [10], TextRank [22], ExpandRank [31] Input documents Statistic-based Wan'07 [32], SC [17], TPR [16] External text corpus, e.g., Wikipedia article Semantics-based Our method Knowledge graphs, e.g., DBpedia.
These processes take a linear time cost proportional to the number of input documents.
