By Lei Wang📧, R. Gopal, R. Shankar, and J. Pancras
In Production and Operations Management, 2022. 31(7), 2773-2788. https://doi.org/10.1111/poms.13727
Customers are increasingly utilizing location-based services via mobile devices to engage with retail establishments. The focus of this paper is to identify factors that help to drive venue popularity revealed by location-based services, which then better facilitate companies’ operational decisions, such as procurement and staff scheduling. Using data collected from Foursquare and Yelp, we build, evaluate, and compare a wide variety of machine learning methods including deep learning models with varying characteristics and degrees of sophistication. First, we find that support vector regression is the best performing model compared to other complex predictive algorithms. Second, we apply SHAP (Shapley Additive exPlanations) to quantify the contribution from each business feature at both the global and local levels. The global interpretability results show that customer loyalty, the agglomeration effect, and the word-of-mouth effect are the top three drivers of venue popularity. Furthermore, the local interpretability analysis reveals that the contributions of business features vary, both quantitatively and directionally. Our findings are robust with respect to different popularity measures, training and testing periods, and prediction horizons. These findings extend our knowledge of location-based services by demonstrating their potential to play a prominent role in attracting consumer engagement and boosting venue popularity. Managers can make better operational decisions such as procurement and staff scheduling based on these more accurate venue popularity prediction methods. Furthermore, this study also highlights the importance of model interpretability which enhances the ability of managers to more effectively utilize machine learning models for effective decision-making.
Keywords: Interpretable machine learning; Location-based services; SHAP (Shapley Additive exPlanations) value; User engagement; Venue popularity prediction
Over the past decade, 
A firm’s success critically hinges on its strategies in selecting its portfolio of products and services. In this paper, we study how differentiation and market alignment at the offering level impact firm performance. To that end, we mine firms’ 10-K filings to characterise the portfolio of offerings through the lens of outward relatedness, inward relatedness, and digitisation. We define outward relatedness as a measure of alignment of firm offerings within its market space, inward relatedness as a measure of differentiation of firm offerings with its own past offerings, and digitisation as a measure of the firm’s focus on IT. We find that markets react positively to firms that operate with high levels of outward relatedness, low levels of inward relatedness and high levels of digitisation. However, we find that highly digitised firms do not have to conform to peers’ offerings. Digitisation enables these firms to differentiate by internally diversifying their offerings. Interestingly, our results show that only firms already highly digitised benefit from further digitisation.
Cryptocurrency price forecasting plays an important role in financial markets. Traditional approaches face two challenges: (1) it is difficult to ascertain the influential factors related to price forecasting; and (2) due to the 24/7 trading policy, cryptocurrencies’ prices face very large fluctuations, thus weakening the forecasting power of traditional models. To address these issues, we focus on Bitcoin and identify the influential factors related to its price forecasting from the perspective of underlying blockchain transactions. We then propose a price forecasting model WT-CATCN, which leverages Wavelet Transform (WT) and Casual Multi-Head Attention (CA) Temporal Convolutional Network (TCN), to forecast cryptocurrency prices. Our model can capture important positions of input sequences and model the correlations among different data features. Using real-world Bitcoin trading data, we test and compare WT-CATCN with other state-of-the-art price forecasting models. The experiment results show that our model improves the price forecasting performance by 25%.
Problem definition: We examine the importance of expert services in online materials and waste exchanges (OMWEs), which are online business-to-business markets for coordinating transactions of industrial surplus, by-products, and waste. Academic/practical relevance: OMWEs face unique challenges because of their product mix and market characteristics. Many OMWEs have traditionally relied on a combination of routine services (online aggregation, filtered search, etc.) and expert services (selective and spatial matching, contract facilitation, etc.). Although OMWEs employ varying levels of expert services, the ultimate value of expert services in promoting transactions is not fully understood. From a managerial perspective, our study provides insights into the importance of balancing routine and expert services, offering guidance on when expert services offer the most benefits. From an academic perspective, we expand on the type of product and market attributes that should be considered in tailoring OMWE designs. Methodology: We use transactional data from a unique OMWE setting (MNExchange.org), which consists of approximately 3,500 product listings from 700+ supplier firms, collected during 2001–2007. We use various econometric techniques (survival analysis, regression discontinuity, etc.) to examine the changes in performance, including transaction rates and time to market, attributable to an operational policy change that occurred in 2004. We further conduct a detailed examination of mechanisms, alternative explanations, and counterfactual analysis. Results: The results show that eliminating expert services in OMWEs can adversely affect transaction outcomes in OMWEs. In particular, the results show that OMWEs should consider their product mix and market characteristics when making decisions about the appropriate use of expert services. Managerial implications: The study provides insights for improving the potential of online reuse marketplaces in the circular economy. From a broader perspective, the paper contributes to the debate on the role of technology in sustainable development and technology substitution for human tasks.
Accurate demand forecasting is critical for supply chain efficiency, especially for the pharmaceutical (pharma) supply chain due to its unique characteristics. However, limited data have prevented forecasters from pursuing advanced models. Such problems exist even when long history of demand data is available because historical data in the distant past may bring little value as market situation changes. In the meantime, demands are also affected by many hidden factors that again require a large amount of data and more sophisticated models to capture. We propose to overcome these challenges by a novel demand forecasting framework which “borrows” time series data from many other products (cross-series training) and trains the data with advanced machine learning models (known for detecting patterns). We further improve performance of the cross-series models through various “grouping” schemes, and learning from non-demand features such as downstream inventory data across different products, information of supply chain structure, and relevant domain knowledge. We test our proposed framework with many modeling possibilities on two large datasets from major pharma manufacturers and our results show superior performance. Our work also provides empirical evidence of the value of downstream inventory information in the context of demand forecasting. We conduct prior and post-hoc field work to ensure the applicability of the proposed forecasting approach.
Recently, diffusion processes in social networks have attracted increasing attention within computer science, marketing science, social sciences, and political science. Although the majority of existing works focus on maximizing the reach of desirable diffusion processes, we are interested in deploying a group of monitors to detect malicious diffusion processes such as the spread of computer worms. In this work, we introduce and study the 
Animal behavior usually has a hierarchical structure and dynamics. Therefore, to understand how the neural system coordinates with behaviors, neuroscientists need a quantitative description of the hierarchical dynamics of different behaviors. However, the recent end-to-end machine-learning-based methods for behavior analysis mostly focus on recognizing behavioral identities on a static timescale or based on limited observations. These approaches usually lose rich dynamic information on cross-scale behaviors. Here, inspired by the natural structure of animal behaviors, we address this challenge by proposing a parallel and multi-layered framework to learn the hierarchical dynamics and generate an objective metric to map the behavior into the feature space. In addition, we characterize the animal 3D kinematics with our low-cost and efficient multi-view 3D animal motion-capture system. Finally, we demonstrate that this framework can monitor spontaneous behavior and automatically identify the behavioral phenotypes of the transgenic animal disease model. The extensive experiment results suggest that our framework has a wide range of applications, including animal disease model phenotyping and the relationships modeling between the neural circuits and behavior.
Gamification has been shown to encourage contributions of user-generated reviews (word-of-mouth, WOM) in various domains, including travel and leisure related platforms (Foursquare, TripAdvisor), e-commerce (Amazon), and auctions (eBay). WOM contributors write reviews about products/services provided by business venues and WOM consumers read reviews and use them to form attitudes and make purchase decisions. Gamification elements such as points and badges, awarded to WOM contributors for various reasons, and displayed to WOM consumers, have a dual role in WOM context. First, points awarded for user contributions help motivate WOM contributors to increase their participation. Second, badges awarded to users for visiting business venues signal prior experience or competence, and they help determine how WOM consumers perceive WOM contributors and form their judgments based on the reviews. While the first role of gamification (i.e., motivating users) has been widely studied, the impact of WOM presented along with gamification elements on the perceptions and behavior of the target audience, WOM consumers, has not been examined. This is important to businesses that are looking to attract customers. Drawing on social psychology literature, we show that gamification symbols signaling experience that accompanies WOM leads to perceptions of positive WOM contributors as more competent. This leads to important changes in behavioral outcomes such as willingness to visit/buy and willingness to recommend the reviewed outlets.