Given a corpus of employee peer reviews from a large corporation where each review is structured into pros and cons, what are the prevalent traits that employees talk about? How can we describe the performance of an employee with just a few sentences, that help us interpret what their work is praised and criticized for? What is the best way to summarize an employee's reviews, while preserving the content and sentiment as well as possible? In this work, we study a large collection of corporation-wide employee peer reviews from a technology enterprise. Motivated by the challenges we outline in our analysis of employee review data, our work makes two main contributions in the domain of people analytics: (a) Sentiment-Aspect Model (SAM): we introduce a stylized log-linear model that identifies the hidden aspects and sentiment within an employee peer review corpus, (b) Interpretable Sentiment-Aspect Representations (EMPLOYEE2VEC): using SAM, we produce a vector space embedding for each employee, containing an overall sentiment score per aspect, and (c) Summarization of Employee Peer Reviews (PEERSUM): we summarize an employee's peer reviews with just a few sentences which reflect the most prevalent traits and associated sentiment for the employee as much as possible. We show that our model SAM can use the structure present in the dataset as supervision to discover meaningful latent traits and sentiment embodied in the reviews. Our employee vector representations Employee2vecprovide a compact, interpretable overview of their evaluation. The review summaries extracted by PeersUmprovide text that explains the professional performance of an employee in a succinct and objectively quantifiable way. We also show how to use our techniques for people analytics tasks such as the analysis of thematic differences between departments, regions, and genders.

We consider the problem of adaptively routing a fleet of cooperative vehicles within a road network in the presence of uncertain and dynamic congestion conditions. To tackle this problem, we first propose a Gaussian process dynamic congestion model that can effectively characterize both the dynamics and the uncertainty of congestion conditions. Our model is efficient and thus facilitates real-time adaptive routing in the face of uncertainty. Using this congestion model, we develop efficient algorithms for non-myopic adaptive routing to minimize the collective travel time of all vehicles in the system. A key property of our approach is the ability to efficiently reason about the long-term value of exploration, which enables collectively balancing the exploration/exploitation trade-off for entire fleets of vehicles. Our approach is validated by traffic data from two large Asian cities. Our congestion model is shown to be effective in modeling dynamic congestion conditions. Our routing algorithms also generate significantly faster routes compared to standard baselines, and achieve near-optimal performance compared to an omniscient routing algorithm. We also present the results from a preliminary field study, which showcases the efficacy of our approach.

We study the problem of large-scale social identity linkage across different social media platforms, which is of critical importance to business intelligence by gaining from social data a deeper understanding and more accurate profiling of users. This paper proposes HYDRA, a solution framework which consists of three key steps: (I) modeling heterogeneous behavior by long-term behavior distribution analysis and multi-resolution temporal information matching; (II) constructing structural consistency graph to measure the high-order structure consistency on users' core social structures across different platforms; and (III) learning the mapping function by multi-objective optimization composed of both the supervised learning on pair-wise ID linkage information and the cross-platform structure consistency maximization. Extensive experiments on 10 million users across seven popular social network platforms demonstrate that HYDRA correctly identifies real user linkage across different platforms, and outperforms existing state-of-the-art algorithms by at least 20% under different settings, and 4 times better in most settings.

Existing algorithms for trajectory-based clustering usually rely on simplex representation and a single proximity-related distance (or similarity) measure. Consequently, additional information markers (e.g., social interactions or the semantics of the spatial layout) are usually ignored, leading to the inability to fully discover the communities in the trajectory database. This is especially true for human-generated trajectories, where additional fine-grained markers (e.g., movement velocity at certain locations, or the sequence of semantic spaces visited) can help capture latent relationships between cluster members. To address this limitation, we propose TODMIS: a general framework for Trajectory cOmmunity Discovery using Multiple Information Sources. TODMIS combines additional information with raw trajectory data and creates multiple similarity metrics. In our proposed approach, we first develop a novel approach for computing semantic level similarity by constructing a Markov Random Walk model from the semantically-labeled trajectory data, and then measuring similarity at the distribution level. In addition, we also extract and compute pair-wise similarity measures related to three additional markers, namely trajectory level spatial alignment (proximity), temporal patterns and multi-scale velocity statistics. Finally, after creating a single similarity metric from the weighted combination of these multiple measures, we apply dense sub-graph detection to discover the set of distinct communities. We evaluated TODMIS extensively using traces of (i) student movement data in a campus, (ii) customer trajectories in a shopping mall, and (iii) city-scale taxi movement data. Experimental results demonstrate that TODMIS correctly and efficiently discovers the real grouping behaviors in these diverse settings.

We consider the problem of adaptively routing a fleet of cooperative vehicles within a road network in the presence of uncertain and dynamic congestion conditions. To tackle this problem, we first propose a Gaussian Process Dynamic Congestion Model that can effectively characterize both the dynamics and the uncertainty of congestion conditions. Our model is efficient and thus facilitates real-time adaptive routing in the face of uncertainty. Using this congestion model, we develop an efficient algorithm for non-myopic adaptive routing to minimize the collective travel time of all vehicles in the system. A key property of our approach is the ability to efficiently reason about the long-term value of exploration, which enables collectively balancing the exploration/exploitation trade-off for entire fleets of vehicles. We validate our approach based on traffic data from two large Asian cities. We show that our congestion model is effective in modeling dynamic congestion conditions. We also show that our routing algorithm generates significantly faster routes compared to standard baselines, and achieves near-optimal performance compared to an omniscient routing algorithm. We also present the results from a preliminary field study, which showcases the efficacy of our approach.

The execution of an agent's complex activities, comprising sequences of simpler actions, sometimes leads to the clash of conflicting functions that must be optimized. These functions represent satisfaction, short-term as well as long-term objectives, costs and individual preferences. The way that these functions are weighted is usually unknown even to the decision maker. But if we were able to understand the individual motivations and compare such motivations among individuals, then we would be able to actively change the environment so as to increase satisfaction and/or improve performance. In this work, we approach the problem of providing highlevel and intelligible descriptions of the motivations of an agent, based on observations of such an agent during the fulfillment of a series of complex activities (called sequential decisions in our work). A novel algorithm for the analysis of observational records is proposed. We also present a methodology that allows researchers to converge towards a summary description of an agent's behaviors, through the minimization of an error measure between the current description and the observed behaviors. This work was validated using not only a synthetic dataset representing the motivations of a passenger in a public transportation network, but also real taxi drivers' behaviors from their trips in an urban network. Our results show that our method is not only useful, but also performs much better than the previous methods, in terms of accuracy, efficiency and scalability.

With the increasing availability of metropolitan transportation data, such as those from vehicle Global Positioning Systems (GPSs) and road-side sensors, it has become viable for authorities, operators, and individuals to analyze the data for better understanding of the transportation system and, possibly, improved utilization and planning of the system. We report our experience in building the Visual Analytics for Intelligent Transportation (VAIT) system, which is the first system on real-life large-scale data sets for intelligent transportation. Our key observation is that metropolitan transportation data are inherently visual as they are spatio-temporal around road networks. Therefore, we visualize and manage traffic data, together with digital maps, and support analytical queries through this interactive visual interface. As a case study, we demonstrate VAIT on real-world taxi GPS and meter data sets from 15 000 taxis running for two months in a Chinese city of over 10 million people. We discuss the technical challenges in data calibration, storage, visualization, and query processing and offer first-hand lessons learned from developing the system. Based on our extensive empirical experiment results, VAIT beats state-of-the-art methods and systems in terms of scalability, efficiency, and effectiveness and offers us an easy-to-use, efficient, and scalable platform to shed more light on intelligent transportation research.

We present an efficient approach to ambulance fleet allocation and dynamic redeployment, where the goal is to position an entire fleet of ambulances to base locations to maximize the service level (or utility) of the Emergency Medical Services (EMS) system. We take a simulation-based approach, where the utility of an allocation is measured by directly simulating emergency requests. In both the static and dynamic settings, this modeling approach leads to an exponentially large action space (with respect to the number of ambulances). Futhermore, the utility of any particular allocation can only be measured via a seemingly “black box” simulator. Despite this complexity, we show that embedding our simulator within a simple and efficient greedy allocation algorithm produces good solutions. We derive data-driven performance guarantees which yield small optimality gap. Given its efficiency, we can repeatedly employ this approach in real-time for dynamic repositioning. We conduct simulation experiments based on real usage data of an EMS system from a large Asian city, and demonstrate significant improvement in the system’s service levels using static allocations and redeployment policies discovered by our approach.

An accurate and sufficient vehicle trajectory data set is the basis to many trajectory-based data mining tasks and applications. However, vehicle trajectories sampled by GPS devices are usually at a relatively low sampling rate and contain notable location errors. To address these two problems in GPS trajectory data, we propose WI-matching, the first vehicle trajectory calibration framework to take advantage of road networks topology and geometry information and trajectory historical information in large scale. WI-matching consists of a Weighting-based map matching algorithm and a trajectory Interpolation-based matching algorithm. In our WI-matching framework, we first integrate the vehicle GPS data with digital road networks data, to identify the roads where a vehicle traveled and the vehicle locations along the roads. Then our weighting-based map matching algorithm considers (1) the geometric and topological information of the road networks and (2) the spatiotemporal trajectory information to efficiently and effectively calibrate the GPS data points. Finally, our interpolation algorithm identifies paths between consecutive GPS points, and adds points with estimated vehicle status (location and time stamp) along the paths to construct sufficient vehicle trajectories. We have evaluated our algorithms on a large-scale real life data set in comparison with the state of the art. Our extensive and empirical results indicate that our WI-matching achieves a high accuracy as well as a high efficiency on real-world data which beats the state of the art.

The global economy and society increasingly depends on computer networks linked together by the Internet. The importance of computer networks reaches far beyond the telecommunications sector since they have become a critical factor for many other crucial infrastructures and markets. With threats mounting and security incidents becoming more frequent, concerns about network security grow. It is an acknowledged fact that some of the most fundamental network protocols that make the Internet work are exposed to serious threats. One of them is the Border Gateway Protocol (BGP) which determines how Internet traffic is routed through the topology of administratively independent networks that the Internet is comprised of. Despite the existence of a steadily growing number of BGP security proposals, to date none of them has been adopted. Using a precise definition of BGP robustness we experimentally show that the degree of robustness is distributed unequally across the administrative domains of the Internet, the so-called Autonomous Systems (ASes). The experiments confirm the intuition that the contribution ASes are able to make towards securing the correct working of the inter-domain routing infrastructure by deploying countermeasures against routing attacks differ depending on their position in the AS topology. We also show that the degree of this asymmetry can be controlled by the choice of the security strategy. We compare the strengths and weaknesses of two fundamentally different approaches in increasing BGP's robustness which we termed ingress and egress detection of false route advertisements and indicate their implications. Our quantitative results have important implications for Internet security policy, in particular with respect to the crucial question where to start the deployment of which type of security scheme in order to maximize the Internet's robustness to routing attacks.

Users of online services are increasingly wary that their activities could disclose confidential information on their business or personal activities. It would be desirable for an online document service to perform text retrieval for users, while protecting the privacy of their activities. In this article, we introduce a privacy-preserving, similarity-based text retrieval scheme that (a) prevents the server from accurately reconstructing the term composition of queries and documents, and (b) anonymizes the search results from unauthorized observers. At the same time, our scheme preserves the relevance-ranking of the search server, and enables accounting of the number of documents that each user opens. The effectiveness of the scheme is verified empirically with two real text corpora.

The World Wide Web has enabled anybody with a low cost Internet connection to access vast information repositories. Some of these repositories contain information (e.g., hate speech and pornography) that is considered objectionable, especially for children to view. Several efforts---legal and technical---are underway to protect children and the generic public from accessing this type of content. We propose a technical approach utilizing a recently proposed technique called isotonic separation for filtering with content metadata if they satisfy monotone conditions. We illustrate this approach using a category rating method of PICS. In essence, we formulate the Internet content filtering problem as a classification problem on content metadata and report on experiments we conducted with the isotonic separation technique.

A version of cobweb/classit is proposed to incrementally cluster text documents into cluster hierarchies. The modification to classit consists of changes to the underlying distributional assumption of the original algorithm that are suggested by text document data. Both the algorithms are evaluated using standard text document datasets. We show that the modified algorithm performs better than the original Classit when presented with Reuters newswire articles in temporal order, i.e., the order in which they are going to be presented in real life situation. It also performs better than the original Classit on the larger of eleven standard text clustering datasets we used.

Summary form only given. Emerging caching techniques such as differential caching, push caching, etc enable preferential treatment of premium objects in a cache. By the use of appropriate placement, replacement or reporting policies, a service provider can offer multiple levels of QoS for different objects. For example, certain objects may be assigned higher priorities and be placed in a premium cache before they get pushed into the regular cache to increase their lifetime in the cache. Similar techniques open avenues for a service provider such as an IAP (Internet Access Provider) to offer several quality-differentiated services to content publishers at different prices and thus maximize revenues. We focus on the optimal pricing strategy of a caching service provider offering multiple quality levels. We analyze trade-offs between bandwidth savings and the potential to charge publishers for caching their content. We determine the optimum number of quality levels to offer and motivate a means for determining optimal allocation of the cache space among the different service classes.

Using data collected between August 1999 and January 2000, we examine firm-level pricing strategies for 32 online bookstores.The data cover more than 300 books, including New York Times bestsellers, computer bestsellers, and random books.Over the period, standard deviation of price fell slightly, and prices for the different types of books were substantially above cost and flat or increasing.Thus predicted price convergence has not yet occurred.An average firm priced random paperback nonfiction books at 94 percent of the cover price and offered substantial discounts (19%-6%) for current and former bestsellers on the New York Times list and the Amazon computer list, small (4%) discounts for hardcover books, and substantial (11%) discounts for fiction books.Significant variation in pricing policies across stores, however, suggests that some firms offer differentiated products and that other firms follow dominated pricing policies.

With the ubiquitous availability of browsers and internet access, the last few years have seen a tremendous growth in the number of applications being developed on the world wide web (WWW). Models for analyzing and designing these applications are only just beginning to emerge. In this work, we propose a 3-dimensional classification space for WWW applications, consisting of a degree of structure of pages dimension, a degree of support for interrelated events dimension and a location of processing dimension. Next, we propose usability design metrics for WWW applications along the structure of pages dimension. To measure these ,we propose CMU-WEB: a conceptual model that can be used to design WWW applications, such that its schemas provide values for the design metrics. This work represents the first effort, to the best of our knowledge, to provide a conceptual model that measures quantifiable metrics that can be used for the design of more usable web applications, and that can also be used to compare the usability of existing web applications, without empirical testing.Request access from your librarian to read this article's full text.

The World Wide Web has already affected OR/MS work in a significant way, and holds great potential for changing the nature of OR/MS products and the OR/MS software economy. Web technologies are relevant to OR/MS work in two ways. First, the Web is a multimedia communication system. Originally based on an information pull model, it is—critically for OR/MS—being extended for information push as well. Second, it is a large distributed computing environment in which OR/MS products—interactive computational applications—can be made available, and interacted with, over a global network. Enabling technologies for Web-based execution of OR/MS applications are classified into those involving client-side execution and server-side execution. Methods for combining multiple client-side and server-side technologies are critical to OR/MS's use of these technologies. These methods, and various emerging technologies for developing computational applications, give the OR/MS worker a rich armament for building Web-based versions of conventional applications. They also enable a new class of distributed applications working on real-time data. Web technologies are expected to encourage the development of OR/MS products as specialized component applications that can be bundled to solve real-world problems. Effective exploitation, for OR/MS purposes, of these technological innovations will also require initiatives, changes, and greater involvement by OR/MS organizations.

We would like to begin by thanking all the commentators for their careful review and thoughtful comments. We see three themes related to OR/MS and the Web in the various comments and suggestions made by the commentators. The first relates to electronic commerce, the second to enabling technology infrastructure, and the third to the role of the Web in serving the INFORMS community. We respond to each of these points in turn.

The use of strong typing, exemplified in the Ascend modeling language, is a recent phenomenon in executable modeling languages for mathematical modeling. It is also one that has significant potential for improving the functionality of computer-based modeling environments. Besides being a strongly typed language, Ascend is unique in providing operators that allow dynamic type inference, a feature that has been shown to be useful in assisting model evolution and reuse. We develop formal semantics for the type system in Ascend—focusing on these operators—and analyze its mathematical and computational properties. We show that despite the strong interactions between various statements involving the operators, the language does possess certain desirable mathematical and computational properties. Further, our analysis identifies general issues in the design and implementation of type systems in mathematical modeling languages. The methods used in the article are applicable beyond Ascend to a class of typed modeling languages that may be developed in the future.

Mathematical modeling systems are an important component in the toolkit of decision support system (DSS) generators. These systems provide a language in which mathematical models can be stated, and enable access to algorithms which may be used to manipulate (e.g., optimize, what-if analysis) the model. While these features are useful, we believe that these systems should also support other important aspects of the DSS model development process. We focus on two such aspects in this paper. These include support for the iterative refinement of models, and for the management of multiple model versions which are created when alternative problem scenarios are investigated. Building on empirical studies of the model development process, we argue that support for these aspects of model development requires knowledge of the rationale for modeling decisions, the dependencies that relate them, and of methods that maintain their consistency. We then demonstrate how belief maintenance concepts developed in the artificial intelligence literature can be used to implement our proposals. While the issues we discuss are relevant to all kinds of mathematical modeling, we examine them in the context of mathematical programming model development and present an implemented system called MODFORM.>

The tomato-tomahto problem (known as the synonymy problem in the database literature) arises in the context of model management when different names are used in different models for what should be identical variables, and these different models are to be integrated or combined into a larger model. When this problem occurs, it is said that the unique names assumption has been violated. We propose a method by which violations of the unique names assumption can be automatically detected. The method relies on declaring four kinds of information and modeling variables: dimensional information, laws relating dimensional expressions, information (called the quiddity) about the intended interpretation of the variables, and laws relating quiddity expressions. We present and discuss the method and the principles and theory behind it, and we describe our (prototype) implementation of the method, as an additional function of an existing model management system. INFORMS Journal on Computing, ISSN 1091-9856, was published as ORSA Journal on Computing from 1989 to 1995 under ISSN 0899-1499.