Customer reviews submitted at Internet travel portals are an important yet underexplored new resource for obtaining feedback on customer experience the hospitality industry. These data often voluminous and unstructured, presenting analytical challenges traditional tools that were designed well-structured, quantitative data. We adapt methods from natural language processing machine learning to illustrate how hotel industry can leverage this source by performing automated evaluation of quality...

Problem description: Measuring quality in the service industry remains a challenge. Existing methodologies are often costly and unscalable. Furthermore, understanding how elements of contribute to performance providers continues be concern industry. In this paper, we address these challenges restaurant sector, vital component Academic/practical relevance: Our work provides scalable methodology for measuring using vast amount text social media. The metrics proposed associated with economic...

Problem definition: Two disciplines increasingly applied in operations management (OM) are machine learning (ML) and behavioral science (BSci). Rather than treating these as mutually exclusive fields, we discuss how they can work complements to solve important OM problems. Methodology/results: We illustrate ML BSci enhance one another non-OM domains before detailing each step of their respective research processes benefit the other settings. then conclude by proposing a framework help...

Time series of graphs are increasingly prevalent in modern data and pose unique challenges to visual exploration pattern extraction. This paper describes the development application matrix factorizations for time-varying community detection time-evolving graph sequences. The factorization model allows user home on display interesting, underlying structure its evolution over time. methods scalable weighted networks with a large number time points or nodes can accommodate sudden changes...

Identifying biomarkers for tuberculosis (TB) is an ongoing challenge in developing immunological correlates of infection outcome and protection. Biomarker discovery also necessary aiding design testing new treatments vaccines. To effectively predict progression any disease, including TB, large amounts experimental data are required to reach statistical power make accurate predictions. We took a two-pronged approach using both computational modeling address this problem. first collected 200...

Problem definition: The effect of the minimum wage is an important yet controversial topic that has received attention for decades. Our study first to take operational lens and empirically impact on firms’ scheduling practices. Methodology/results: Using a highly granular data set from chain fashion retail stores, we estimate $1 increase in wage, although having negligible total labor hours used by leads 27.7% number workers scheduled per week, but 19.4% reduction weekly worker. For average...

AbstractThree-dimensional data arrays (collections of individual matrices) are increasingly prevalent in modern and pose unique challenges to pattern extraction visualization. This article introduces a biclustering technique for exploration detection such complex structured data. The proposed framework couples the popular plaid model together with tools from functional analysis guide estimation bicluster responses over array. We present an efficient algorithm that first detects biclusters...

From an upset stomach to a life-threatening foodborne illness, getting sick is all too common after eating in restaurants. Although health inspection programs for restaurants are designed protect consumers, such inspections typically occur sporadically, allowing restaurant hygiene remain unknown diners. At the same time, online reviews provide valuable source of information about current status and quality restaurant. In this paper, we use text contained these effectively identify cases...

We study the behavior of interbank market before, during and after 2008 financial crisis. Leveraging recent advances in network analysis, we two structures, a correlation based on publicly traded bank returns, physical lending transactions. While networks behave similarly pre-crisis, crisis shows an increase interconnectedness while highlights marked decrease interconnectedness. Moreover, these respond differently to monetary macroeconomic shocks. Physical forecast liquidity problems crises.

We study the behavior of interbank market before, during and after 2008 financial crisis. Leveraging recent advances in network analysis, we two structures, a correlation based on publicly traded bank returns, physical lending transactions. While networks behave similarly pre-crisis, crisis shows an increase interconnectedness while highlights marked decrease interconnectedness. Moreover, these respond differently to monetary macroeconomic shocks. Physical forecast liquidity problems crises.

Not long ago securities were traded by human traders in face-to-face markets. The ecosystem of an open outcry market was well-known, visible to a eye, and rigidly prescribed. Now trading is increasingly done anonymous electronic markets where do not have designated functions or mandatory roles. In fact, the themselves been replaced algorithms (machines) operating with little no oversight. While process machine-learning methods developed recognize persistent patterns data. this study, we...

Grade inflation is pervasive in educational settings the United States. One driver of grade may be faculty concern that assigning lower clinical performance scores to trainees will cause them retaliate and assign teaching member. The finding near-zero retaliation would important members who evaluate trainees.The authors used a bidirectional confidential evaluation feedback system test hypothesis residents subsequently receive scores. From September 1, 2008, February 15, 2013, 177 evaluated...

In this article we create a novel monitoring system to detect changes within sequence of networks. Specifically, consider sparse, weighted, directed, and attributed Our approach uses the Hurdle model capture sparsity explain weights edges as function node edge attributes. Here, weight an represents number interactions between two nodes. We then integrate with state-space temporal dynamics formation process. Estimation is performed using extended Kalman Filter. Statistical process control...

Data extracted from social media platforms are both large in scale and complex nature, since they contain unstructured text, as well structured data, such time stamps interactions between users. A key question for is to determine influential users, the sense that generate members of platform. Common measures used academic literature by companies provide analytics services variants popular web-search PageRank algorithm applied networks capture connections In this work, we develop a modeling...

Mobile apps are one of the building blocks mobile digital economy. A differentiating feature to traditional enterprise software is online reviews, which available on app marketplaces and represent a valuable source consumer feedback app. We create supervised topic modeling approach for developers use reviews as useful sources quality customer feedback, thereby complementing testing. The based constrained matrix factorization that leverages relationship between term frequency given response...

The rise of social media platforms has fundamentally altered the public discourse by providing easy to use and ubiquitous forums for exchange ideas opinions. Elected officials often such communication with broader disseminate information engage their constituencies other officials. In this work, we investigate whether Twitter conversations between legislators reveal real-world position influence analyzing multiple networks that feature different types link relations Members Parliament (MPs)...

Identifying critical components in networked systems is a key problem for many important applications diverse set of fields, including epidemiology, e-commerce and traffic systems. This paper describes the development application semi-nonnegative matrix factorization structural discovery featuring nodes that are transmission over social networks. The technique allows practitioner to perform structured by specifying context-specific network statistics guide solution. techniques demonstrated...

This study investigates two-stage plans based on nonparametric procedures for estimating an inverse regression function at a given point. Specifically, isotonic is used stage one to obtain initial estimate followed by another round of in the vicinity this two. It shown that such accelerate convergence rate one-stage and are superior existing use local parametric approximations two when available budget moderate and/or “ill-behaved.” Both Wald- likelihood ratio-type confidence intervals...

Problem Definition: Consider a network of stores operating under the same brand, for example, chain coffee shops or banks. Increasing sales at one store may have different impact on another store: from negative as result potential cannibalization to positive increased brand awareness and customer engagement. We study how causally identify spatially heterogeneous effects between pairs, which can be used measure influence.Academic/Practical Relevance: Our work provides scalable methodology...

Online reviews and their effect on business outcomes have long been of interest to information systems scholars. In this study, we complement the existing research online by proposing a novel use modern text analysis methods uncover semantic structure assess impact survival merchants in marketplace. We analyze from 2005 2014 for restaurants major metropolitan area United States find that variables capturing within review are important predictors restaurants, relationship has not explored...