outcome; an accurate classifier should have an AUC of more than 0.50 ( Fawcett, 2006 ). These values were calculated using Hugin Researcher's analysis wizard. The predictive accuracy of the candidate BBN models was compared with logistic regression
Arthur Villordon, Ron Sheffield, Jose Rojas, and Yin-Lin Chiu
Arthur Villordon, Craig Roussel, and Tad Hardy
The sweetpotato weevil [SPW, Cylas formicarius (Fabricius)] is an important economic pest in “pink-tagged” or SPW-infested areas of Louisiana. From time to time, sweetpotato weevils are detected in “green-tagged” or SPW-free locations. When sweetpotato weevils are detected in “green'tagged” areas, the produce is quarantined and may not be shipped to locations that do not allow “pink-tagged” sweetpotatoes. As part of the statewide SPW monitoring program, the Louisiana Department of Agriculture and Forestry (LDAF) conducts a statewide pheromone-based trapping program to monitor SPW presence in beds and fields. We used SPW presence-absence data with a GIS-based logistic regression modeling tool to assess the feasibility of developing a model for predicting SPW risk in sweetpotato beds. Using pheromone trap data from 2001–03, we performed stepwise logistic regression experiments to assess the role of various weather variables (daily mean maximum and minimum temperature, rainfall) in the occurrence of SPW in beds. Our modeling experiments showed a strong relationship of mean daily minimum temperature during the winter months with SPW occurrence in beds. In particular, a logistic regression equation developed from 2003 trap data and mean April daily minimum temperature created a spatially accurate map of SPW risk for 2002. However, the same model did not accurately predict the 2001 SPW risk. These results indicate that additional variables are needed to improve the predictive ability of the model. Spatial risk mapping can be a potentially useful tool for decision makers in choosing between risk-averse and -prone decisions.
Anna K. Kirk and Rufus Isaacs
To maximize yield of pollination-dependent agricultural crops, farmers must ensure that sufficient pollinators are present when flowers are open and viable. We characterized and compared the lower development threshold temperature, bloom phenology, and flower viability of five common cultivars of highbush blueberry (Vaccinium corymbosum L.) to enable prediction of when flowers would be available for pollination. Threshold temperatures of all cultivars were found to be very similar and range between 7 and 8 °C. Logistic regression was used to characterize bloom phenology for all cultivars under field and greenhouse conditions. Bloom phenology under greenhouse conditions was delayed ≈100 growing degree-days when compared with field conditions. Average flower viability was determined daily from first flower opening until 5 days after flower opening for each cultivar. Results indicated declining flower viability with increasing flower age with most flowers unsuitable for pollination more than 4 days after opening. Implications of these results for planning pollination of highbush blueberry fields are discussed.
Mwamburi Mcharo*, Don Labonte, Chris Clark, and Mary Hoy
Using two sweetpotato (Ipomoea batatas (L.) Lam) F1 populations from diverse environments we investigated the AFLP marker profiles of the genotypes for association studies between the molecular markers and southern root-knot nematode (Meloidogyne incognita) resistance expression. Population one consisted of 51 half-sib genotypes developed at the Louisiana State Univ. AgCenter. The second population consisted of 51 full-sibs developed by the East African and International Potato Center sweetpotato breeding programs. Results for nematode resistance expression indicate a binomial distribution among the genotypes. Using analysis of molecular variance, logistic regression and discriminant analysis, AFLP markers that are most influential with respect to the phenotypic trait expression were selected for both populations. A comparative analysis of the power of models from the two statistical models for southern root-knot nematode resistance class prediction was also done. The diversity and possible universal similarity of influential markers between the two populations and the expected impact in sweetpotato breeding programs will be discussed.
Scott Kalberer, Norma Leyva-Estrada, Stephen Krebs, and Rajeev Arora
Winter survival of temperate-zone woody perennials requires them to resist loss of frost hardiness (deacclimation) during winter and early spring thaws. However, little is known about deacclimation response in woody landscape plants. Moreover, what impact, if any, the degree of deacclimation has on reacclimation capacity has not been systematically studied. We used nine genotypes of deciduous azaleas (Rhododendron subgenus Pentanthera) to investigate effects of deacclimating conditions on bud cold hardiness and reacclimation ability. Dormant floral buds, with 3–5 cm stem attached, were collected in late December from field-grown plants, and placed in constant warm [22 °C 15 °C (D/N)] and humid conditions for increasing durations (0-day to 14-day) to stimulate deacclimation. Bud cold hardiness (lt 50) was determined (using logistic regressions) by evaluating immature flower survival at subfreezing treatment temperatures. Results indicated that azalea genotypes from colder provenances showed greater initial frost hardiness. Typically northern genotypes had slow to intermediate deacclimation rates, while rates of southern genotypes were intermediate to rapid. High initial frost hardiness was frequently associated with slow deacclimation. Buds retained the capacity to reacclimate upon cold exposure [2 °C/–2 °C; (12 h/12 h)] even after 8 days of deacclimation. Distinct differences were observed between the two latitudinal ecotypes of R. viscosum with respect to their initial bud hardiness, deacclimation rates, and reacclimation capacities. We suggest that the three attributes, i.e., high initial hardiness, slow deacclimation, and high reacclimation capacity, together may be important for winter-survival of azalea buds.
Julie M. Tarara, Paul E. Blom, Bahman Shafii, William J. Price, and Mercy A. Olmstead
Estimates of canopy and fruit fresh mass are useful for more accurate interpretation of data from the Trellis Tension Monitor, a tool for real-time monitoring of plant growth and predicting yield in trellised crops. In grapevines (Vitis labruscana Bailey), measurements of shoot and fruit fresh mass were collected at frequent intervals (14 to 21 days) over 5 years, and these data were correlated with variables that could be obtained nondestructively: shoot length, number of leaves per shoot, and number of clusters per shoot. Shoot length provided a good estimator of shoot fresh mass in all years. Nonlinear logistic regression models described the dynamics of canopy growth from bloom to the early stages of ripening, which often is poorly represented by simple linear regression approaches to seasonal data. A generalized function indicated a lower bound of ≈600 degree-days, after which an increase in shoot fresh mass could be considered on average to contribute only slightly to further increases in trellis wire tension. The dynamics of fruit mass were captured adequately by a nonlinear function, but not as well as vegetative mass because of larger variances in fruit mass. The number of clusters per shoot was associated with fruit mass only after the accumulation of ≈550 degree-days or, equivalently, the time at which fruit mass exceeded ≈25 g per shoot. Seasonal dynamics of the ratio of fruit to vegetative fresh mass were not sufficiently discernable by the logistic models because of the dominance of fruit mass and its large interannual variation.
Product behavior represents how consumers perceive and use a product. Its importance in predicting consumer buying behavior is well documented in marketing research. There are, however, no data available investigating the role of product behavior in the floral market. This study addressed this deficiency. Data were first analyzed using factor analysis to extract the principal determinants of product behavior in the floral market. As a result, six primary behavioral factors were identified and named as: “using flowers as daily essentials,” “perceived product value,” “negative attitude toward flowers,” “using flowers as gifts,” “eventbased usage,” and “experience in receiving flowers.” The effects of these extracted behavioral factors on consumer flower purchase frequency were then further investigated with multinomial logistic regression analysis. Analytical results revealed that behaviors “using flowers as daily essentials” and “using flowers as gifts” forced consumers to become heavy users in the floral market. Conversely, “negative attitude toward flowers” negatively affected the floral purchase frequency. Experimental results in this study also suggest that promoting a positive attitude toward flowers is essential in encouraging consumers to become flower users. The intended use of flower product purchase, whether for personal use or as gifts, was the main factor affecting the frequent purchasing of flowers.
Patricia Garriz, Hugo Alvarez, and Graciela Colavita
Temperature has long been recognized as a major environmental factor affecting the net carbon exchange in the pear tree, as well as the growth of fruit. The objective of this work was to predict pear fruit growth as a function of accumulated growing-degree-days (DD) using a mathematical model. A crop of `Abbé Fetel' trees was studied at the Experimental Farm of the Comahue National University, Argentina (lat. 38°56'S; long. 67°59'W). Maximum fruit diameter (FD) measurements were carried out every 2 weeks during three growing seasons (2000–01, 2001–-02, and 2002–03). The range of sampling dates was 26 and 143 days after full bloom (DFB). An automated meteorological station, situated close to the orchard, collected temperature data, which were expressed on the basis of DD from time of full bloom to harvest, with critical temperatures at 4 and 35 °C. Equations were developed with SYSTAT procedure and model suitability was evaluated using goodness-to-fit measures. It was found that the following logistic regression provided the most satisfactory fit for the pooled data: FD (mm) = 71.62/[(1 + e^(1.7450-0.0027DD)], coefficient of determination = 0.96. The testing on an independent crop showed that predictions were accurate. Analyses of fruit growth, based on DD, did not improve data interpretation over that on a DFB basis. The average monthly temperature varied little between seasons. A remarkably consistent heat-unit accumulation function was obtained from year to year, with a 5% maximum variation in number of DFB to maturity, compared to a 6% variation in DD, occurring between the 2002 and 2003 commercial harvests. These results have important implications for cultural practices, such as fruit thinning and final size forecast.
Bryan J. Peterson, Gregory J.R. Melcher, Ailish K. Scott, Rebecca A. Tkacs, and Andrew J. Chase
cuttings that rooted was assessed by logistic regression in JMP 14.1.0 (SAS Institute, Cary, NC) because the influence of independent variables on the probability of binary responses is not suited to analysis of variance (ANOVA). Fisher’s exact test was
Arancha Arbeloa, Ma Elena Daorden, Elena García, Pilar Andreu, and Juan A. Marín
) Germination rate versus embryo size in the F1 (pooled data). The logistic regression coefficient (LRC) that is analogous to the “slope” in conventional linear regression analysis and the explained deviance (ED) by the model expressed as the percentage of total