Comparative Document Analysis for Large Text Corpora

Abstract

This paper presents a novel research problem on joint dis

Assessment and prioritization of the WHO “best buys” and other recommended interventions for the prevention and control of non-communicable diseases in Iran

“…This is a mixed-method study. We used a comparative framework [19][20][21], to conduct document analysis for comparing the existing policies about NCDs in Iran with the WHO's recommended interventions and policies on NCDs. MCDA quantitative approach was also used to do priority setting of preventive intervention in two areas: reducing modifiable risk factors and strengthening and reorientation of the health systems to address the prevention and control of NCDs.…”

“…Also, the senior author (AT) is a member of INCDC and had frequent and ongoing interactions with WHO, i.e., global meetings on NCDs as well as informal contacts with staff and representatives of many countries, from local to global levels. To design a comparative framework [19][20][21], we considered objectives and interventions for each of these objectives as a basis for comparison and matching (Table 2). Finally, we collected national policy and documents (see Appendix A) related to the identified objectives, actions, and interventions, from the MoHME and scrutinized them in the same way.…”

Foundations of Temporal Text Networks

“…In [24] vertices represent either famous characters from the text or bags of words, while the edges connect words that best explain the contexts where two or more famous characters appear together in the text. Document-phrase graphs as defined in [23] are also HIN-based models, and more in detail probabilistic bipartite networks B = (V, U, E, W ) where the vertices in one partition V represent documents from a large document collection, the vertices in U represent salient phrases which are semantically relevant to one or more documents in V , and edges E indicate the relevance of each sentence for each document. HINs are not limited to represent relations within documents, text and concepts; but they can also model relations between actors and text.…”

Phrase Mining from Massive Text and Its Applications

“…en the ontology-like topological order ensures these content units have the chance of being jointly activated by general phrase nodes via inter-phrase links. Many techniques [Dahab et al, 2008, Sanderson and Croft, 1999, Yin and Shah, 2010 have been previously developed to induce an ontological structure over quality phrases. It is out of scope of our work to specifically address these or evaluate their relative impact in our evaluation.…”

“…Sentence-based summarization, on the other hand, may be too verbose to highlight the general commonalities and differences-users may be distracted by the irrelevant information contained there. Recent studies [Ren et al, 2017a leverage quality phrases, i.e., minimal semantic unit, to summarize the commonalities and differences. the relation between document subsets induced by query context and identify phrases that truly distinguish the queried subset of documents from neighboring subsets.…”