Ying Lü

Decision tree methodology is a commonly used data mining method for establishing classification systems based on multiple covariates or for developing prediction algorithms for a target variable. This method classifies a population into branch-like segments that construct an inverted tree with a root node, internal nodes, and leaf nodes. The algorithm is non-parametric and can efficiently deal with large, complicated datasets without imposing a complicated parametric structure. When the sample size is large enough, study data can be divided into training and validation datasets. Using the training dataset to build a decision tree model and a validation dataset to decide on the appropriate tree size needed to achieve the optimal final model. This paper introduces frequently used algorithms used to develop decision trees (including CART, C4.5, CHAID, and QUEST) and describes the SPSS and SAS programs that can be used to visualize tree structure.

UNLABELLED: We measured cortical and trabecular bone loss using QCT of the spine and hip in 14 crewmembers making 4- to 6-month flights on the International Space Station. There was no compartment-specific loss of bone in the spine. Cortical bone mineral loss in the hip occurred primarily by endocortical thinning. INTRODUCTION: In an earlier study, areal BMD (aBMD) measurements by DXA showed that cosmonauts making flights of 4- to 12-month duration on the Soviet/Russian MIR spacecraft lost bone at an average rate of 1%/month from the spine and 1.5%/month from the hip. However, because DXA measurements represent the sum of the cortical and trabecular compartments, there is no direct information on how these bone envelopes are affected by spaceflight. MATERIALS AND METHODS: To address this, we performed a study of crewmembers (13 males and 1 female; age range, 40-55 years) on long-duration missions (4-6 months) on the International Space Station (ISS). We used DXA to obtain aBMD of the hip and spine and volumetric QCT (vQCT) to assess integral, cortical, and trabecular volumetric BMD (vBMD) in the hip and spine. In the heel, DXA was used to measure aBMD, and quantitative ultrasound (QUS) was used to measure speed of sound (SOS) and broadband ultrasound attenuation (BUA). RESULTS AND CONCLUSIONS: aBMD was lost at rates of 0.9%/month at the spine (p < 0.001) and 1.4-1.5%/month at the hip (p < 0.001). Spinal integral vBMD was lost at a rate of 0.9%/month (p < 0.001), and trabecular vBMD was lost at 0.7%/month (p < 0.05). In contrast to earlier reports, these changes were generalized across the vertebrae and not focused in the posterior elements. In the hip, integral, cortical, and trabecular vBMD was lost at rates of 1.2-1.5%/month (p < 0.0001), 0.4-0.5%/month (p < 0.01), and 2.2-2.7%/month (p < 0.001), respectively. The cortical bone loss in the hip occurred primarily by cortical thinning. Calcaneal aBMD measurements by DXA showed smaller mean losses (0.4%/month) than hip or spine measurements, with SOS and BUA showing no change. In summary, our results show that ISS crewmembers, on average, experience substantial loss of both trabecular and cortical bone in the hip and somewhat smaller losses in the spine. These results do not support the use of calcaneal aBMD or QUS measurements as surrogate measures to estimate changes in the central skeleton.