Applied Stochastic Models in Business and Industry最新文献

North star metrics and online experimentation play a central role in how technology companies improve their products. In many practical settings, however, evaluating experiments based on the north star metric directly can be difficult. The two most significant issues are (1) low sensitivity of the north star metric and (2) differences between the short-term and long-term impact on it. A common solution is to rely on proxy metrics rather than the north star in experiment evaluation and launch decisions. Existing literature on proxy metrics concentrates mainly on the estimation of the long-term impact from short-term experimental data. In this article, instead, we focus on the trade-off between the estimation of the long-term impact and the sensitivity in the short term. In particular, we propose the Pareto optimal proxy metrics method, which simultaneously optimizes prediction accuracy and sensitivity. We also give a multi-objective optimization algorithm to solve our specific problem. We apply our methodology to experiments from a large industrial recommendation system, and found proxy metrics that are eight times more sensitive than the north star and consistently moved in the same direction, increasing the velocity and the quality of the decisions to launch new features.

The paper arises from the experience of Applied Stochastic Models in Business and Industry which has seen, over the years, more and more contributions related to Machine Learning rather than to what was intended as a stochastic model. The very notion of a stochastic model (e.g., a Gaussian process or a Dynamic Linear Model) can be subject to change: What is a Deep Neural Network if not a stochastic model? The paper presents the views, supported by examples, of distinguished researchers in the field of business and industrial statistics. They are discussing not only whether there is a future for traditional stochastic models in the era of Machine Learning and Artificial Intelligence, but also how these fields can interact and gain new life for their development.

The aim of this paper is to investigate the problem of one and two active redundant components allocation in a k-out-of-n system with dependent components. Here, some necessary and sufficient conditions are presented under which the redundancies are optimally allocated to the system components based on the usual stochastic order criterion. In addition, it is shown that, unlike the independence mode, a redundant component is not necessarily allocated to the weakest component. Further, in the case of the two redundant components, the weak (strong) redundant component is not necessarily allocated to the stronger (weaker) component of the system. Some algorithms are also presented for calculating the reliability of the considered system under the assumption of dependency between the main and redundant components. Using different copula functions for describing the dependencies between components, various examples are given to illustrate the optimal allocation of redundant components.

Cryptocurrencies exhibit high volatility, emphasizing the importance of accurately measuring tail risk in their markets. This research incorporates a threshold-switching mechanism into Taylor's ES-CAViaR models that unveil features such as asymmetry and jump phenomena. These enhancements effectively capture the diverse tail risks of cryptocurrencies while enabling the simultaneous forecasting of both Value-at-Risk (VaR) and Expected Shortfall (ES). The proposed models incorporate two types of functions to address the VaR and ES nexus with the option to use the rolling standard deviation of returns as a short-term volatility proxy as a regressor. We estimate the parameters and forecast tail risk within a Bayesian framework. Taking the two largest cryptocurrencies by market capitalization, Bitcoin and Ethereum, we assess the one-step-ahead forecasting performance over a four-year out-of-sample period using a rolling window approach. The comparative results from backtests and five scoring functions among eight competing models support the conclusion that models with a threshold mechanism capture the tail risk of cryptocurrencies more accurately than other risk models.

The fourth industrial revolution has driven the emergence of Digital Twins (DTs) and Industrial Internet of Things (IIoT) in manufacturing. However, the use of different definition has led to varied interpretations and inconsistent understanding of DTs. Thus, by exploring the gap between theoretical frameworks and practical implementations of IIoT-based DTs in manufacturing, this paper aims to shed light on the DT phenomenon by considering the historical evolution and fundamental concepts of IIoT-based DTs. Therefore, a systematic literature review was conducted to assess the ambiguity concerning DTs, particularly in distinguishing architectures and types. Therefore, this paper identifies IIoT-based DTs in manufacturing by reviewing application-oriented literature. As a result of a subsequent classification, this paper proposes a hierarchical classification based on communication dynamics (i.e., Uni-directional and Bi-directional) and information processing (i.e., use or non-use of machine learning). Conclusively, this study proposes a comprehensive classification approach for IIoT-based DTs and thus contributes to a more consistent understanding of the DT phenomenon. Moreover, this paper discusses key findings, as well as implications for research and practice. Finally potential avenues for future research are derived and the limitations of this study are discussed.

A flexible model for analysing load-sharing data is developed by approximating the cumulative hazard functions of component lifetimes by piecewise linear functions. The proposed model is data-driven and does not depend on restrictive parametric assumptions on underlying component lifetimes. Maximum likelihood estimation and construction of confidence intervals for model parameters are discussed. Estimates of reliability characteristics such as reliability at a mission time, quantile function, mean time to failure and mean residual time for load-sharing systems are developed in this setting. As the proposed model is capable of providing a good fit for load-sharing data, it also results in a better estimation of these important reliability characteristics. The performance of the proposed model is observed to be quite satisfactory through a detailed Monte Carlo simulation study. The analyses of two load-sharing datasets, one pertaining to the lives of two-motor load-sharing systems and another related to basketball games, are provided as illustrative examples. In summary, this article presents a comprehensive discussion on a flexible model that can be used for load-sharing systems efficiently.