Comparison Between Nelson-Olson Method and Two-Stage Limited Dependent Variables (2SLDV ) Method for the Estimation of a Simultaneous Equations System (Tobit Model)

This study relates to the estimation of a simultaneous equations system for the Tobit model where the dependent variables ( ) are limited, and this will affect the method to choose the good estimator. So, we will use new estimations methods different from the classical methods, which if used in such a case, will produce biased and inconsistent estimators which is (Nelson-Olson) method and Two- Stage limited dependent variables(2SLDV) method to get of estimators that hold characteristics the good estimator . That is , parameters will be estimated for the limited variables and find the variance-covariance matrix for extracted estimators by the aforementioned two methods and then compare between the results of the two methods and find any better method by estimation and then finding the estimation efficiency, and this is what the study aims to . A simultaneous equations system will be imposed for the limited model defined by two equations containing two endogenous variables one of complete observations and the other censored at zero. The two methods were used to analyze the relationship between income and family expenditure on durable consumer goods , where the results showed that the performance of (Nelson-Olson) method is better than performing the Two-Stage limited dependent variables (2SLDV) method in obtain the lower values and all comparison measures as well as the results showed that income and expenditure one affects the other and the and the price affects the income and expenditure

Measuring spatial price differentials at the basic heading level: a comparison of stochastic index number methods

Spatial price comparisons rely to a high degree on the quality of the underlying price data that are collected within or across countries. Below the basic heading level, these price data often exhibit large gaps. Therefore, stochastic index number methods like the Country–Product–Dummy (CPD) method and the Gini–Eltetö–Köves–Szulc (GEKS) method are utilised for the aggregation of the price data into higher-level indices. Although the two index number methods produce differing price level estimates when prices are missing, the present paper demonstrates that both can be derived from exactly the same stochastic model. For a specific case of missing prices, it is shown that the formula underlying these price level estimates differs between the two methods only in weighting. The impact of missing prices on the efficiency of the price level estimates is analysed in two simulation studies. It can be shown that the CPD method slightly outperforms the GEKS method. Using micro data of Germany’s Consumer Price Index, it can be observed that more narrowly defined products improve estimation efficiency.

INCREASING THE ESTIMATION EFFICIENCY OF THE FUNCTIONAL CHARACTERISTICS OF GRAIN HARVESTERS

When assessing the functional characteristics of combine harvesters, one of the most important indicators to be determined is grain loss occ urring in the threshing-and-separating unit. In compliance with GOST 28301-2015 “Combine harvesters. Test methods”, it should not exceed 1.5%. The existing methods of its determination and estimation are laborious and require specialized frames of samplers manually placed in the grain in front of the combine or the use of strip samplers. Moreover, technologies for grain harvesting with chopping and spreading straw require the use of an automated system for placing sample frames. The authors present an automatic device for placing sampling frames under the threshing-and-separating unit of a grain harvester during its operation. The proposed design fi ve sampling frames with a size of 1500×650 mm both in manual and automatic modes with a predetermined interval. The device provides for better labor safety of testers, reduces the complexity of testing by 10…15%, and improves the accuracy of the agrotechnical and operational-technological assessment of grain harvesters.

Exploiting Distributional Temporal Difference Learning to Deal with Tail Risk

In traditional Reinforcement Learning (RL), agents learn to optimize actions in a dynamic context based on recursive estimation of expected values. We show that this form of machine learning fails when rewards (returns) are affected by tail risk, i.e., leptokurtosis. Here, we adapt a recent extension of RL, called distributional RL (disRL), and introduce estimation efficiency, while properly adjusting for differential impact of outliers on the two terms of the RL prediction error in the updating equations. We show that the resulting “efficient distributional RL” (e-disRL) learns much faster, and is robust once it settles on a policy. Our paper also provides a brief, nontechnical overview of machine learning, focusing on RL.

Use of Remotely Sensed Data to Enhance Estimation of Aboveground Biomass for the Dry Afromontane Forest in South-Central Ethiopia

Periodic assessment of forest aboveground biomass (AGB) is essential to regulate the impacts of the changing climate. However, AGB estimation using field-based sample survey (FBSS) has limited precision due to cost and accessibility constraints. Fortunately, remote sensing technologies assist to improve AGB estimation precisions. Thus, this study assessed the role of remotely sensed (RS) data in improving the precision of AGB estimation in an Afromontane forest in south-central Ethiopia. The research objectives were to identify RS variables that are useful for estimating AGB and evaluate the extent of improvement in the precision of the remote sensing-assisted AGB estimates beyond the precision of a pure FBSS. Reference AGB data for model calibration and estimation were collected from 111 systematically distributed circular sample plots (SPs) of 1000 m2 area. Independent variables were derived from Landsat-8, Sentinel-2 and PlanetScope images acquired in January 2019. The area-weighted mean and standard deviation of the spectral reflectance, spectral index and texture (only for PlanetScope) variables were extracted for each SP. A maximum of two independent variables from each image type was fitted to a generalized linear model for AGB estimation using model-assisted estimators. The results of this study revealed that the Landsat-8 model with the predictor variable of shortwave infrared band reflectance and the PlanetScope model with the predictor variable of green band reflectance had estimation efficiency of 1.40 and 1.37, respectively. Similarly, the Sentinel-2 model, which had predictor variables of shortwave infrared reflectance and standard deviation of green leaf index, improved AGB estimation with the relative efficiency of 1.68. Utilizing freely available Sentinel-2 data seems to enhance the AGB estimation efficiency and reduce cost and extensive fieldwork in inaccessible areas.

Estimation efficiency of rewound induction motors in situ using a numerical model

This paper presents an effective technique for determining the impact of rewinding practices on the motor efficiency and characterizing the efficiency reduction when electrical motors are rewound several times. This technique focuses on a new approach and a statistical study to find a numerical model for the estimation efficiency of rewound induction motors in the field. The experimental results from 101 induction motor tests are analyzed. A numerical model is determined and compared with different methods: separate losses method, modified current method and simple current method. An error analysis is conducted to examine the level of uncertainty by testing three asynchronous motors at 110 kW, 160 kW, and 300 kW. The results show that this approach can predict and estimate the efficiency reduction in rewound motors without expensive tests and can help the energy manager make effective cost decisions in replacing the rewound motors with more efficient ones by using an assessment of overconsumption and maintenance costs. Another advantage of this model is that it can be used to improve the software tools and can also be a very strong indicator to audit the repair quality

A working likelihood approach for robust regression

Robust approach is often desirable in presence of outliers for more efficient parameter estimation. However, the choice of the regularization parameter value impacts the efficiency of the parameter estimators. To maximize the estimation efficiency, we construct a likelihood function for simultaneously estimating the regression parameters and the tuning parameter. The “working” likelihood function is deemed as a vehicle for efficient regression parameter estimation, because we do not assume the data are generated from this likelihood function. The proposed method can effectively find a value of the regularization parameter based on the extent of contamination in the data. We carry out extensive simulation studies in a variety of cases to investigate the performance of the proposed method. The simulation results show that the efficiency can be enhanced as much as 40% when the data follow a heavy-tailed distribution, and reaches as high as 468% for the heteroscedastic variance cases compared to the traditional Huber’s method with a fixed regularization parameter. For illustration, we also analyzed two datasets: one from a diabetics study and the other from a mortality study.

PERFORMANCE EVALUATION OF ENERGY RECONSTRUCTION METHODS IN HIGH ENERGY PHYSICS EXPERIMENTS

The discovery of particles that shape our universe pushes the scientific community to increasingly build sophisticated equipment. Particle accelerators are one of these complex machines that put known particle beams on a collision course at speeds close to that of light. When collisions occur, subproducts are produced and measured by the calorimeter system, which entirely absorbs these subproducts. Typically, a high-energy calorimeter is highly segmented, comprising thousands of dedicated readout channels. The present work evaluates the performance of two energy reconstruction algorithms: the OF (Optimal Filter) and MAE (Multi-Amplitude Estimator), which was recently proposed to deal with the signal superposition (pile-up). In order to evaluate the energy estimation efficiency, artificial data were used, considering several pile-up levels. The statistics from the energy estimation is employed to compare the performance achieved by each method. A second analysis is made to quantify the MAE sensitivity to the pedestal parameter. The results show that the MAE method presents a better performance than the OF method and the usage of an uncalibrated pedestal value compromises the MAE performance.