It is important for Bayesian network (BN) structure learning, a NP-problem, to improve the accuracy and hybrid algorithms are a kind of effective structure learning algorithms at present. Most hybrid algorithms adopt ...
详细信息
It is important for Bayesian network (BN) structure learning, a NP-problem, to improve the accuracy and hybrid algorithms are a kind of effective structure learning algorithms at present. Most hybrid algorithms adopt the strategy of one heuristic search and can be divided into two groups: one heuristic search based on initial BN skeleton and one heuristic search based on initial solutions. The former often fails to guarantee globality of the optimal structure and the latter fails to get the optimal solution because of large search space. In this paper, an efficient hybrid algorithm is proposed with the strategy of two-stage searches. For first-stage search, it firstly determines the local search space based on Maximal Information Coefficient by introducing penalty factors p(1), p(2), then searches the local space by Binary Particle Swarm Optimization. For second-stage search, an efficient adr (the abbreviation of Add, Delete, Reverse) algorithm based on three basic operators is designed to extend the local space to the whole space. Experiment results show that the proposed algorithm can obtain better performance of BN structure learning.
Critically ill newborns in neonatal intensive care units (NICUs) are at greater risk of developing adverse drug reactions (adrs). Differentiation of adrs from reactions associated with organ dysfunction/immaturity is ...
详细信息
Critically ill newborns in neonatal intensive care units (NICUs) are at greater risk of developing adverse drug reactions (adrs). Differentiation of adrs from reactions associated with organ dysfunction/immaturity is difficult. Current adr algorithm scoring was established arbitrarily without validation in infants. The study objective was to develop a valid and reliable algorithm to identify adrs in the NICU. algorithm development began with a 24-item questionnaire for data collection on 100 previously suspected adrs. Five pediatric pharmacologists independently rated cases as definite, probable, possible, and unlikely adrs. Consensus gold standard was reached via teleconference. Logistic regression and iterative C programs were used to derive the scoring system. For validation, 50 prospectively collected adr cases were assessed by 3 clinicians using the new algorithm and the Naranjo algorithm. Weighted kappa and intraclass correlation coefficient (ICC) were used to compare validity and reliability of algorithms. The new algorithm consists of 13 items. Kappa and ICC of the new algorithm were 0.76 and 0.62 versus 0.31 and 0.43 for the Naranjo algorithm. The new algorithm developed using actual patient data is more valid and reliable than the Naranjo algorithm for identifying adrs in the NICU population. Because of the relatively small and nonrandom samples, further refinement and additional testing are needed.
暂无评论