Bayesian optimization of multivariate genomic prediction models based on secondary traits for improved accuracy gains and phenotyping costs

Key messageWe propose a novel approach to the Bayesian optimization of multi-variate genomic prediction models based on secondary traits to improve accuracy gains and phenotyping costs via efficient Pareto frontier estimation.Multivariate genomic prediction based on secondary traits, such as data from various omics technologies including high-throughput phenotyping (e.g., unmanned aerial vehicle-based remote sensing), has attracted much attention because it offers improved accuracy gains compared with genomic prediction based only on marker genotypes. Although there is a trade-off between accuracy gains and phenotyping costs of secondary traits, no attempt has been made to optimize these trade-offs. In this study, we propose a novel approach to optimize multivariate genomic prediction models for secondary traits measurable at early growth stages for improved accuracy gains and phenotyping costs. The proposed approach employs Bayesian optimization for efficient Pareto frontier estimation, representing the maximum accuracy at a given cost. The proposed approach successfully estimated the optimal secondary trait combinations across a range of costs while providing genomic predictions for only about 20% of all possible combinations. The simulation results reflecting the characteristics of each scenario of the simulated target traits showed that the obtained optimal combinations were reasonable. Analysis of real-time target trait data showed that the proposed multivariate genomic prediction model had significantly superior accuracy compared to the univariate genomic prediction model.

Assessing the market power of millers and wholesalers in the Bangladesh rice sector

PurposeBecause of the increasing differential between farm and retail prices, the study proposes to investigate the extent of market power in the rice value chain of Bangladesh using advanced econometric techniques.Design/methodology/approachUsing a Stochastic Frontier Estimation approach on cross-sectional data, the study examines the price spread along the rice value chain to determine whether millers and wholesalers exercise market power.FindingsEmpirical results reveal that, on average, rice millers and wholesalers charge 33 and 29% above the marginal cost, respectively. This study confirms the non-competitive behavior of the rice market with wholesalers and millers wielding substantial market powerResearch limitations/implicationsA limitation of the study is that it does not include the retailers who also play a major role in the Bangladesh rice value chain. This is left for future study.Originality/valueThis study combines primary and secondary data collected on the Bangladesh rice sector to examine the market power of two major players along the value chain, millers and wholesalers, using an advanced econometrics approach.

Efficiency in the Public University

Usual efficiency evaluation methods, such as Data Envelopment Analysis and Stochastic Frontier Estimation tend to calculate efficiency scores which are valid only at a given moment and for the units where in the sample. At most, Malmquist index determines efficiency scores for period t+1. Those drawbacks can be avoided by translating efficiency scores in terms of probabilities trough the Discriminant Analysis which enables also, to provide appropriate weightings activities of research laboratories, object of this study. In addition to Discriminant Analysis, Hierarchical Cluster Analysis is used to classify those activities into homogeneous groups, while Data Envelopment Analysis is employed to categorize laboratories in efficient and inefficient structures. This method is applied on a public university in east of Morocco. This study shows that indexed publications as well as the supported doctoral thesis or in state of supervision have the greatest impact on efficiency of the research laboratories.

Dynamic frontier estimation for monitoring team performances

Purpose This paper aims to investigate how predictions of football league standings and efficiency measures of teams, obtained through frontier estimation technique, evolve compared to actual results. Design/methodology/approach The study is based on data from the Turkish first division football league. Historical data for five seasons, from 2011 to 2016, are used to compare weekly estimates to de facto results. Data envelopment analysis efficiency measures are used to estimate team performances. After each week, a data envelopment analysis is run using available data until then, and final team standings are estimated via computed efficiencies. Estimations are improved by using a data envelopment analysis model that incorporates expert knowledge about football. Findings Results indicate that deductions can be made about the league’s future progress. Model incorporating expert knowledge tends to estimate the performance better. Although the prediction accuracy starts out low in early stages, it improves as the season advances. Scatter of individual teams’ performances show fluxional behaviour, which attracts studying the impact of uncontrollable factors such as refereeing. Originality/value While all previous studies focus on season performance, this study handles the problem as a combination of weekly performance and how it converges to reality. By tracking weekly performance, managers get a chance to confront their weak performance indicators and achieve higher ranking by improving on these inefficiencies.