Citation explanations

1. A system for providing citation explanations comprising:

an extraction engine to extract a citation claim from a primary document using a syntactic analysis of a portion of the primary document;

a segment generation engine to identify a set of candidate segments of a cited document that may correspond to the citation claim; and

an evaluation engine to analyze the set of candidate segments to rank each candidate segment of the set of candidate segments based on an explainability score calculated by the system using a combination of a candidate segment position factor, a content similarity value, a structural class factor, and an information value factor, wherein the candidate segment position factor varies based on whether a candidate segment position is in a first or last paragraph of a document or is in a first or last sentence of a paragraph, wherein the system is to display a candidate segment based on the explainability score.

2. The system of claim 1 , wherein the extraction engine performs a pre-processing of the citation claim to provide a representation of the citation claim.

3. The system of claim 1 , wherein the extraction engine classifies the citation claim based on assignment of predefined classes.

4. The system of claim 1 , wherein the evaluation engine merges at least two of the candidate segments to form a revised candidate segment and provides an updated rank to the revised candidate segment.

5. A non-transitory computer-readable storage medium encoded with instructions that, when executed by a processor performs a method, causes the processor to:

extract and classify a citation claim from a primary document using a syntactic analysis of a portion of the primary document;

generate a set of candidate segments of a cited document that may correspond to the citation claim;

analyze the set of candidate segments to rank each candidate segment of the set of candidate segments based on an explainability score calculated by the method using a combination of a candidate segment position factor, a content similarity value, a structural class factor, and an information value factor, wherein the candidate segment position factor varies based on whether a candidate segment position is in a first or last paragraph of a document or is in a first or last sentence of a paragraph; and

display a candidate segment based on the explainability score.

6. The non-transitory computer-readable storage medium of claim 5 , wherein the citation claim comprises at least one structure selected from a vector of words, a topic representation, and a set of exact words.

7. The non-transitory computer-readable storage medium of claim 5 , wherein the candidate segment of the set of candidate segments is selected for display based on the explainability score.

8. A method to provide citation explanations comprising:

analyzing a primary document to extract a citation claim;

identifying a set of candidate segments of a cited document that may correspond to the citation claim by comparing the cited document and the citation claim to generate at least one candidate segment that matches or explains the citation claim;

calculating an explainability score;

evaluating the set of candidate segments to rank each candidate segment of the set of candidate segments based on the explainability score calculated using a combination of a candidate segment position factor, a content similarity value, a structural class factor, and an information value factor, wherein the candidate segment position factor varies based on whether a candidate segment position is in a first or last paragraph of a document or is in a first or last sentence of a paragraph; and

displaying a candidate segment based on the explainability score.

9. The method of claim 8 , further comprising analyzing the primary document using a syntactic analysis that analyzes of the primary document based on at least one of the following methods: sentence boundaries, fixed window context of words around the citation, and predefined number of sentences surrounding a citation.

10. The method of claim 8 , further comprising analyzing the cited document to provide a structure for the cited document.

11. The method of claim 8 , further comprising analyzing the cited document to generate an index of the cited document.

12. The method of claim 8 , further comprising selecting a portion of the cited document to represent at least one candidate segment.

13. The method of claim 8 , further comprising merging at least two segments of the set of candidate segments to form a revised candidate segment.

14. The method of claim 13 , further comprising providing an explainability score of the revised candidate segment.