Using k-nn and discriminant analyses to classify rain forest types in a Landsat TM image over northern Costa Rica

Conservation and land use planning in humid tropical lowland forests urgently need accurate remote sensing techniques to distinguish among floristically different forest types. We investigated the degree to which floristically and structurally defined Costa Rican lowland rain forest types can be accurately discriminated by a non-parametric k nearest neighbors (k-nn) classifier or linear discriminant analysis. Pixel values of Landsat Thematic Mapper (TM) image and Shuttle Radar Topography Mission (SRTM) elevation model extracted from segments or from 5 × 5 pixel windows were employed in the classifications. 104 field plots were classified into three floristic and one structural type of forest (regrowth forest). Three floristically defined forest types were formed through clustering the old-growth forest plots (n = 52) by their species specific importance values. An error assessment of the image classification was conducted via cross-validation and error matrices, and overall percent accuracy and Kappa scores were used as measures of accuracy. Image classification of the four forest types did not adequately distinguish two old-growth forest classes, so they were merged into a single forest class. The resulting three forest classes were most accurately classified by the k-nn classifier using segmented image data (overall accuracy 91%). The second best method, with respect to accuracy, was the k-nn with 5 × 5 pixel windows data (89% accuracy), followed by the canonical discriminant analysis using the 5 × 5 pixel window data (86%) and the segment data (82%). We conclude the k-nn classifier can accurately distinguish floristically and structurally different rain forest types. The classification accuracies were higher for the k-nn classifier than for the canonical discriminant analysis, but the differences in Kappa scores were not statistically significant. The segmentation did not increase classification accuracy in this study.     

Computerized and written questionnaires: Are they equivalent?

In this set of three studies, comparisons were made between computer-administered questionnaires and those administered in written form. Specifically, we examined nonclinical subjects' affective reactions to the two modes of questionnaire administration and their responses to a variety of commonly used attitude and personality measures. Subjects' affective reactions to the two modes of administration did not differ. With respect to questionnaire responses, there is some evidence suggesting that computer administration may slightly increase socially desirable responding amongst those subjects with relatively little computer experience. Overall, however, the differences between the two conditions were negligible, suggesting that social scientists should not hesitate to take advantage of the benefits of computer-administered questionnaires.     

Relations between prenatal testosterone and cerebral lateralization in children.

Several theorists have proposed that the sex steroid testosterone acts on the fetal brain during a critical period of development to influence cerebral lateralization (N. Geschwind & A. M. Galaburda, 1987; M. Hines & C. Shipley, see PA, Vol 71:8996; S. F. Witelson, see PA, Vol 79:26441. In the present study. relations were examined between prenatal testosterone levels in 2nd trimester amniotic fluid and lateralization of speech, affect, and handedness at age 10. Girls with higher prenatal testosterone levels were more strongly right-handed and had stronger left-hemisphere speech representation. Boys with higher prenatal testosterone levels had stronger right-hemisphere specialization for the recognition of emotion. This pattern of results is most consistent with Witelson's (1991) claim that prenatal testosterone leads to greater lateralization of function. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Mediation of entry of human immunodeficiency virus-1 into alveolar macrophages by CD4 without facilitation by surfactant-associated protein-A

The mechanism of uptake of human immunodeficiency virus-1 (HIV-1) into alveolar macrophages (AM), freshly isolated blood monocytes (MN), and cultured MN (CM) was investigated focusing on the role of CD4 and of surfactant-associated protein A (SP-A). By radioimmunoassay which obviated the problems of auto- and nonspecific fluorescence of more differentiated macrophages, each of the macrophage populations studied expressed CD4. Semiquantitative polymerase chain reaction was performed to assess uptake of HIV-1(JR-FL) into cells. OKT4a (directed against CD4) blocked uptake of HIV-1 into CM, AM, and MN by 67 to 100%. OKT4 (directed against another epitope of CD4) had a smaller and less consistent effect (0-90%), and control antibodies showed minimal effects and only at supersaturating concentrations. SP-A had no effect on uptake of HIV-1 into AM. SP-A also had no consistent effect on production of HIV-1(JR-FL) by AM infected in vitro (p24 antigen ELISA). Thus CD4 is the major receptor for HIV-1 in mononuclear phagocytes, including AM, and SP-A does not modulate entry.     

Relations between prenatal testosterone levels and cognitive abilities at 4 years.

Relations between prenatal testosterone (T) levels and cognitive abilities at age 4 were examined for 28 girls and 30 boys. Prenatal T levels were measured in 2nd trimester amniotic fluid samples obtained by amniocentesis and were examined in relation to scores on tests of cognitive abilities. For girls, prenatal T levels showed a curvilinear (inverted U-shaped) relation to language comprehension and classification abilities. Linear relations also were observed in that prenatal T levels were inversely related to girls' scores on tasks assessing counting and number facts. Similarly, girls with high average block building scores had lower prenatal T and cognitive abilities were not observed. The observation of relations in girls and not boys is discussed, and the findings are examined in relation to theories of hormone-behavior relations. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Study of behavioral phenotypes: goals and methodological considerations

The test principle and the optimization of the reactive ingredients are described for the one-step dip and-read immunochromatographic FRONTLINE rapid tests for drugs-of-abuse testing in urine samples. In a multicenter evaluation the rapid tests were compared with FPIA and EMIT immunoassays. Discrepant results were further analyzed by gas chromatography-mass spectrometry methods. In the comparison of the cannabinoids rapid tests versus both immunoassays using clinical and forensic urine samples (399 versus FPIA and 755 versus EMIT), sensitivities and specificities were 97% or better for both comparisons. For cocaine, a sensitivity of 100% versus both routine technologies was obtained, whereas the specificity was reduced somewhat to 91% because of some cross-reactivity with metabolites of methadone and of clozapine. Specificity was very high for the cocaine rapid tests (98-100%) when applied to urine samples of persons not in a methadone maintenance program. Sensitivities and specificities for the opiates rapid tests were 99% or better at all sites when compared with the routine methods. In the screening of about 1200 clinical urine samples for cannabinoids, cocaine or opiates misuse only six samples would have stayed undetected by rapid test analyzes. These results show the FRONTLINE assays allow a reliable and immediate screening for drugs of abuse.     

A study of the thermomechanical properties of carbon fiber-polypropylene composites

Short fiber composites were produced using polypropylene as matrix and polyacrylonitrile (PAN)-based fibers or vapor grown carbon fibers (VGCF) as reinforcement. The strength, stiffness, and coefficient of thermal expansion (CTE) of the composites were measured. The VGCF-composites showed strength and CTE that are competitive with those of “conventional” PAN-fiber composites, but the stiffness was marginally lower. Micromechanical modeling of the PAN composite properties gives results consistent with the measurements. The models can be used to infer the apparent VGCF-properties from their composites.     

Application of rare-earth doped fibres as lowpass filters in passive optical networks

In the passive optical networks currently under study, it is necessary to provide high attenuation of the 1500-1600 nm fibre window in front of the telephony receiver and transmitter. A cheap, simple filter is required which also has low loss at 1300 nm. Fibres doped with suitable concentrations of the rare-earth ions, erbium and neodymium are shown to have spectral attenuation characteristics ideal for this application.<>     

The multispectral separability of Costa Rican rainforest types with support vector machines and Random Forest decision trees

Estimating the extent of tropical rainforest types is needed for biodiversity assessment and carbon accounting. In this study, we used statistical comparisons to determine the ability of Landsat Thematic Mapper (TM) bands and spectral vegetation indices to discriminate composition and structural types. A total of 144 old-growth forest plots established in northern Costa Rica were categorized via cluster analysis and ordination. Locations for palm swamps, forest regrowth and tree plantations were also acquired, making 11 forest types for separability analysis. Forest types classified using support vector machines (SVM), a theoretically superior method for solving complex classification problems, were compared with the random forest decision tree classifier (RF). Separability comparisons demonstrate that spectral data are sensitive to differences among forest types when tree species and structural similarity is low. SVM class accuracy was 66.6% for all forest types, minimally higher than the RF classifier (65.3%). TM bands and the Normalized Difference Vegetation Index (NDVI) combined with digital elevation data notably increased accuracies for SVM (84.3%) and RF (86.7%) classifiers. Rainforest types discriminated here are typically limited to one or two categories for remote sensing classifications. Our results indicate that TM bands and ancillary data combined via machine learning algorithms can yield accurate and ecologically meaningful rainforest classifications important to national and international forest monitoring protocols.     

Visible and near-infrared calibrations for quality assessment of fresh phase I and II mushroom (Agaricus bisporus) compost

Previous studies have shown that visible and near-infrared spectra (Vis-NIR) of dry and milled compost can be used for generating partial least squares (PLS) calibrations of phase II compost parameters including ammonia, nitrogen dry matter (NDM), dry matter (DM), pH, conductivity, carbon, microbial population, and potential productivity. The objective of this study was to develop robust calibrations for some of the key parameters from the spectra of fresh phase I and II composts. Samples of substrates from six commercial production yards were obtained during winter and summer months of 2000-2004 to monitor changes in quality and were analyzed for the test factors. Vis-NIR reflectance measurements of fresh samples (740) were made over the range of 400-2500 nm. After mathematical pretreatments, PLS calibrations of the key parameters were developed using the NIR (1100-2500 nm) and visible and NIR (400-2500 nm) regions and subsequently validated using an independent sample set of 123 phase I and II samples obtained during 2004-2005. The phase I and II standard errors of laboratory measurements of ammonia, pH, conductivity, DM, NDM, and ash were lower than the standard error of predictions of the same parameters, respectively, by the best NIR or Vis-NIR models. The degree of precision for some of the calibrations, especially ammonia, NDM, and DM, is suitable for composters to monitor changes in quality parameters during production. The laboratory measurement errors for phase I samples were greater than those of the phase II samples, except for ash, due to a higher degree of heterogeneity in the substrate. The calibrations, especially for pH, conductivity, and ash, need to be improved with new sample sets. A major advantage of NIR spectroscopy is the ability to assess substrate quality for a range of target parameters simultaneously, within a few hours of receiving the samples. The main drawbacks are the expensive instrumentation, expertise, and training necessary for operating the spectrometer and a dedicated chemometrician required for maintaining the equations compared to the reference methods.