Active learning for prediction of tensile properties for material extrusion additive manufacturing.

Machine learning techniques were used to predict tensile properties of material extrusion-based additively manufactured parts made with Technomelt PA 6910, a hot melt adhesive. An adaptive data generation technique, specifically an active learning process based on the Gaussian process regression algorithm, was employed to enable prediction with limited training data. After three rounds of data collection, machine learning models based on linear regression, ridge regression, Gaussian process regression, and K-nearest neighbors were tasked with predicting properties for the test dataset, which consisted of parts fabricated with five processing parameters chosen using a random number generator. Overall, linear regression and ridge regression successfully predicted output parameters, with < 10% error for 56% of predictions. K-nearest neighbors performed worse than linear regression and ridge regression, with < 10% error for 32% of predictions and 10-20% error for 60% of predictions. While Gaussian process regression performed with the lowest accuracy (< 10% error for 32% of prediction cases and 10-20% error for 40% of predictions), it benefited most from the adaptive data generation technique. This work demonstrates that machine learning models using adaptive data generation techniques can efficiently predict properties of additively manufactured structures with limited training data.

Predictive factors for re-displacement in diaphyseal forearm fractures in children-role of radiographic indices.

Background and purpose - Manipulation and cast immobilization is the primary management for diaphyseal forearm fractures in children, and re-displacement is the most common complication. We wanted (1) to analyze the incidence of re-displacement in a group of children treated with close reduction and casting; (2) to determine predictive factors such as demographics, mechanism of injury, affected bone, fracture pattern, degree of initial displacement and angulation, and reduction accuracy; and (3) to determine the prognostic effect of previously defined radiographic indices. Patients and methods - We prospectively studied 269 consecutive children with closed and complete middle-third diaphyseal fractures treated with close reduction and casting from October 2014 to April 2015. Factors analyzed included demographics, initial fracture features, having a non-anatomical reduction, and the radiographic indices of cast quality. Results - There were 189 fractures of both bones (70%) and 80 solitary fractures (30%). The overall re-displacement rate was 11%. According to multivariable analysis, independent predictors of re-displacement were initial angulation >10° (RR =5) and failure to achieve an anatomical reduction (RR =2). Statistically significant radiographic indices regarding increased rate of re-displacement included cast index ≥0.7 (RR =5), Canterbury index ≥1.1 (RR =3), and 3-point index ≥0.8 (RR =6). Interpretation - Our results suggested that fractures with a higher degree of initial angulation and non-anatomical reduction more often result in re-displacement. Moreover, the casting quality examined with the radiographic indices played an important role in the success of a non-operative management.

Serum gamma-glutamyltransferase, alanine aminotransferase, and aspartate aminotransferase activity in Iranian healthy blood donor men.

AIM: To determine serum gamma-glutamyltransferase (GGT), alanine aminotransferase (ALT), and aspartate aminotransferase (AST) activity, and to assess their correlation with demographic and clinical findings in healthy blood donors. METHODS: This cross-sectional study was performed in 934 male blood donors, aged 18 to 68 years, who consecutively attended Tehran blood transfusion service in 2006. All participants were seronegative for HBV or HCV infections, non alcohol users, and all underwent a standard interview and anthropometric tests. Clinical and biochemical parameters including AST, ALT, and GGT activities were determined. Patients taking drugs known to cause hepatic fat deposition were excluded. For AST, ALT, and GGT variables, we used 33.33 and 66.66 percentiles, so that each of them was divided into three tertiles. RESULTS: Mean AST, ALT, and GGT activities were 25.26 +/- 12.58 U/L (normal range 5-35 U/L), 33.13 +/- 22.98 (normal range 5-35 U/L), and 25.11 +/- 18.32 (normal range 6-37 U/L), respectively. By univariate analyses, there were significant associations between increasing AST, ALT, or GGT tertiles and age, body weight, body mass index, and waist and hip circumferences (P<0.05). By multiple linear regression analyses, ALT was found to be positively correlated with dyslipidemia (B=6.988, P=0.038), whereas ALT and AST were negatively correlated with age. AST, ALT, and GGT levels had positive correlation with family history of liver disease (B=15.763, P<0.001), (B=32.345, P<0.001), (B=24.415, P<0.001), respectively. CONCLUSION: Although we did not determine the cutoffs of the upper normal limits for AST, ALT, and GGT levels, we would suggest screening asymptomatic patients with dyslipidemia and also subjects with a family history of liver disease.