Search

Current research on Web search has focused on optimizing and evaluating single queries. However, a significant fraction of user queries are part of more complex tasks [Jones and Klinkner 2008] which span multiple queries across one or more search sessions [Liu and Belkin 2010; Kotov et al. 2011]. An ideal search engine would not only retrieve relevant results for a user's particular query but also be able to identify when the user is engaged in amore complex task and aid the user in completing that task [Morris et al. 2008; Agichtein et al. 2012]. Toward optimizing whole-session or task relevance, we characterize and address the problem of intrinsic diversity (ID) in retrieval [Radlinski et al. 2009], a type of complex task that requires multiple interactions with current search engines. Unlike existing work on extrinsic diversity [Carbonell and Goldstein 1998; Zhai et al. 2003; Chen and Karger 2006] that deals with ambiguity in intent across multiple users, ID queries often have little ambiguity in intent but seek content covering a variety of aspects on a shared theme. In such scenarios, the underlying needs are typically exploratory, comparative, or breadth-oriented in nature. We identify and address three key problems for ID retrieval: identifying authentic examples of ID tasks from post-hoc analysis of behavioral signals in search logs; learning to identify initiator queries that mark the start of an ID search task; and given an initiator query, predicting which content to prefetch and rank. [ABSTRACT FROM AUTHOR]

Background In the last 20 years, Brazil has undergone dramatic changes in terms of socioeconomic development and health care. In the first decade of the 2000s, the Ministry of Health (MoH) developed a series of programs focused on reducing infant mortality, including the Family Health Program as a national policy for primary care. In this paper, we propose a method to correct underreporting of deaths and live births. After vital statistics are corrected, infant mortality trends are analyzed for the period 2000-2010 by macro-geographical region. Methods A proactive search of live births and deaths was carried out in the Amazon and Northeast regions in 2010 to find vital events that occurred in 2008 and were not reported to the Ministry of Health. The probabilistic sample of 133 municipalities was stratified by adequacy of vital information reporting. For each municipality, the adequacy analysis was based on the reported age-standardized mortality rate per 1,000 population and the ratio between reported and estimated live births. Correction factors were estimated by strata based on additional vital events found in the proactive search. The procedure was generalized to correct municipal vital statistics for the period 2000-2010. Results In the proactive search, 35% of non-reported deaths were found within the health system (hospitals and other health establishments), but 28% were found in non-official sources, like illegal cemeteries. In areas of extreme poverty and unreliable vital information, the estimated completeness of infant death reporting was only 33%. After correction of vital information, the estimated infant mortality rate decreased from 26.1 in 2000 to 16.0 in 2010, with an annual rate of decrease of 4.7%, greater than the required rate to achieve the Millennium Development Goal. Among Brazilian regions, the Northeast showed the largest decrease, from 38.4 to 20.1 per 1,000 live births. Conclusions The proactive search for vital events was shown to be a good strategy both in terms of understanding local irregularities and for correcting vital statistics. The methodology could be applied in other countries to routinely assess the pattern and extent of birth and death under-registration in order to improve the utility of these data to inform health policies. [ABSTRACT FROM AUTHOR]

Abstract: Processing complex and non-routine tasks, engineers face frequent knowledge shortage. While an expert system requires much effort to develop and a general search engine is blamed for passiveness in meeting people’s knowledge demand, we propose in this paper a context-sensitive knowledge supply method which aims at meeting users’ unuttered knowledge need bred in computer-aided engineering (CAE) tasks. To this end, concepts involved in a task are extracted and used for perceiving various problematic situations which may occur; and keyword-based search and text parsing techniques are combined to retrieve possible remedies from unstructured knowledge carriers. The proposed method is tested in situation of finite element analysis (FEA), a typical CAE task, where novice engineers receive sentential knowledge recommendations extracted from webpage. Experiment results show that context-sensitive knowledge supply can increase an engineer’s knowledge about the current task and make the individual more prepared for future challenges. [Copyright &y& Elsevier]