The Shapley value is a fundamental concept in data science, providing a principled framework for fair resource allocation, feature importance quantification, and improved interpretability of complex models. Its fundamental theory is based on four axiomatic proper ties, which underpin its widespread application. To address the inherent computational challenges of exact calculation, we discuss model-agnostic approximation techniques, such as Random Order Value, Least Squares Value, and Multilinear Extension Sampling, as well as specialized fast algorithms for linear, tree-based, and deep learning models. Recent extensions, such as Distributional Shapley and Weighted Shapley, have broadened the applications to data valuation, reinforcement learning, feature interaction analysis, and multi-party cooperation. Practical effectiveness has been demonstrated in health care, finance, industry, and the digital economy, with promising future directions for incorporating these techniques into emerging fields, such as data asset pricing and trading.

As the economic environment becomes increasingly complex, enhancing supply chain resilience is crucial for the operations and long-term development of enterprises. Real-world supply chains, encompassing components such as goods, warehouses, and plants, often contain complex network structures, making resilience analysis a challenging task. This paper addresses this challenge from a network analysis perspective. We project the complex supply chain network into single-mode, multi-layer networks focusing on plants and warehouses. Utilizing a multi-layer community detection method, we identify local clusters within these networks. By uncovering closely connected clusters, we reveal the flexibility and redundancy in production capabilities among different plants and warehouses. An empirical study using real-world data demonstrates that multi-layer network clustering effectively uncovers indirect capacity linkages between plants and warehouses. The findings from this community detection are beneficial for strategic capacity management, aiding enterprises in managing supply shortages or sudden demand spikes.

The emergence of massive data in recent years brings challenges to automatic statistical inference. This is particularly true if the data are too numerous to be read into memory as a whole. Accordingly, new sampling techniques are needed to sample data from a hard drive. In this paper, we propose a sequential addressing subsampling (SAS) method that can sample data directly from the hard drive. The newly proposed SAS method is time saving in terms of addressing cost compared to that of the random addressing subsampling (RAS) method. Estimators (e.g., the sample mean) based on the SAS subsamples are constructed, and their properties are studied. We conduct a series of simulation studies to verify the finite sample performance of the proposed SAS estimators. The time cost is also compared between the SAS and RAS methods. An analysis of the airline data is presented for illustration purpose.

Link prediction is one of the most important personalized services in social network platforms. The key point is to predict the probability of the existence of a link between two nodes based on various information in the network. This article combines information of the network structure with the user-generated contents. We propose link prediction indices based on both network structure and topic distribution (NSTD). In contrast to previous literatures, this approach makes full use of the network characteristics, such as homophily, transitivity, clustering, and degree heterogeneity. And we combine these characteristics with topic similarity when constructing indices based on both directly and indirectly connected nodes. Experiment results demonstrate that the proposed method outperforms the previous methods.