Chetwynd A G, Diggle P J, Drewett R F, Young B
Department of Mathematics and Statistics, Fylde College, Lancaster University, UK.
Stat Med. 1998 Feb 28;17(4):395-405. doi: 10.1002/(sici)1097-0258(19980228)17:4<395::aid-sim778>3.0.co;2-0.
Previous studies of sucking patterns have mainly been on bottle-fed babies and have assumed that the babies' sucks occur within bursts separated by gaps of predetermined minimum length which is fixed over the feed. This study considers babies that are breast-fed, a more complex and natural process than bottle-feeding, and develops a more sophisticated model for the pattern of bursts and gaps which allows the parameters of the process to vary over the feed. We consider data from four breast feeds of each of 32 babies. We develop a two-component mixture model based on an underlying Markov chain model for the switching between bursts and gaps. We use the model to provide summary statistics for each feed and give estimates of the normal range of each of the model's parameters.
以往对吸吮模式的研究主要针对奶瓶喂养的婴儿，并且假定婴儿的吸吮发生在由预定的最小长度间隙分隔的阵发期内，该最小长度在喂养过程中是固定的。本研究考虑母乳喂养的婴儿，这是一个比奶瓶喂养更复杂、更自然的过程，并针对阵发期和间隙模式开发了一个更复杂的模型，该模型允许过程参数在喂养过程中变化。我们考虑了32名婴儿中每名婴儿4次母乳喂养的数据。我们基于一个潜在的马尔可夫链模型开发了一个双组分混合模型，用于阵发期和间隙之间的转换。我们使用该模型为每次喂养提供汇总统计信息，并给出模型每个参数正常范围的估计值。
