Ways to Measure Loyalty of Service Users

The focus of this study was set on the empirical determination of the differences between the methods of measuring the loyalty of users of services. The research reviewed approaches to evaluating the loyalty of taxi service users in Ho Chi Minh (Vietnam), Moscow (Russia), and Novosibirsk (Russia). The goal implied the segmentation of service users depending on the patterns of conduct and giving advice on the development of loyalty programs. The novelty lies in grouping respondents by age and gender and determining the ways to measure their loyalty. The research findings imply that people under 30 tend to make a choice based on reviews of others. The study developed possible customer behavior models and options for actions depending on their satisfaction and loyalty. This matrix can be used for building a comprehensive strategy to increase the loyalty of their clients. Practical value of the study lies in the possibility of using the data obtained to model the behavior of consumers of different service types based on the proposed loyalty matrix.

Empirical determination of sustainable withdrawal rates considering historical yields and inflation rates in Germany

On account of the current low interest rate phase, which is most likely to continue in the coming years, the average yields to be achieved in the bond, time deposit and savings product sectors are declining, so that risk-averse investors in particular have few opportunities to generate return-oriented retirement provisions.This scientific article analyzes the level of a possible safe withdrawal rate for diversified pension portfolios, considering historical returns and inflation rates. Consequently, this article provides immediate practical added value for a possible retirement provision.The evaluation is based on the consideration of historical returns of the stock and bond market in Germany. To determine a safe withdrawal rate, the development of portfolios with different compositions and inflation-adjusted withdrawal rates are simulated over periods of 15 to 35 years. In this simulation, the risky part of the portfolio is represented by German equities, the low risk part by German government bonds.To sum up, the empirical results show a maximum safe withdrawal rate of 4%. The underlying portfolio is composed of 50% equities and 50% government bonds. Particularly due to the outlined demographic change in Germany as well as the ongoing low-interest phase, the empirical study can provide significant theoretical and practical insights.

Fast, accurate and non-empirical determination of the lattice thermal conductivities of I-III-VI2 chalcopyrite semiconductors

The use of computer simulation to predict the lattice thermal conductivity of materials has the potential to accelerate the discovery of new thermoelectric materials. However, the accurate prediction of this property from first principles, without input from experiment, is very computationally demanding, which limits the use of high-throughput strategies in thermoelectric materials design. We present here an accurate, fast, and non-empirical determination of the lattice thermal conductivities of a large family of semiconductors, with composition ABX2 (I-III-VI2), with A=Cu, Ag; B=Al, Ga, In, Tl; and X=S, Se, Te. We solve the Boltzmann transport equation with force constants derived from density functional theory calculations and machine-learning-based regression algorithms, reducing between one and two orders of magnitude the computational cost with respect to conventional approaches of the same accuracy. The results are in good agreement with available experimental data and allow us to rationalize the role of chemical composition, temperature and nanostructuring on the thermal conductivities across this important family of semiconductors.

Combine-Net: An Improved Filter Pruning Algorithm

The powerful performance of deep learning is evident to all. With the deepening of research, neural networks have become more complex and not easily generalized to resource-constrained devices. The emergence of a series of model compression algorithms makes artificial intelligence on edge possible. Among them, structured model pruning is widely utilized because of its versatility. Structured pruning prunes the neural network itself and discards some relatively unimportant structures to compress the model’s size. However, in the previous pruning work, problems such as evaluation errors of networks, empirical determination of pruning rate, and low retraining efficiency remain. Therefore, we propose an accurate, objective, and efficient pruning algorithm—Combine-Net, introducing Adaptive BN to eliminate evaluation errors, the Kneedle algorithm to determine the pruning rate objectively, and knowledge distillation to improve the efficiency of retraining. Results show that, without precision loss, Combine-Net achieves 95% parameter compression and 83% computation compression on VGG16 on CIFAR10, 71% of parameter compression and 41% computation compression on ResNet50 on CIFAR100. Experiments on different datasets and models have proved that Combine-Net can efficiently compress the neural network’s parameters and computation.