ISeeU: Visually interpretable deep learning for mortality prediction inside the ICU
ISeeU: یادگیری عمیق قابل تفسیر برای پیش بینی مرگ و میر در بخش مراقبت های ویژه-2019
To improve the performance of Intensive Care Units (ICUs), the field of bio-statistics has developed scores which try to predict the likelihood of negative outcomes. These help evaluate the effectiveness of treatments and clinical practice, and also help to identify patients with unexpected outcomes. However, they have been shown by several studies to offer sub-optimal performance. Alternatively, Deep Learning offers state of the art capabilities in certain prediction tasks and research suggests deep neural networks are able to outperform traditional techniques. Nevertheless, a main impediment for the adoption of Deep Learning in healthcare is its reduced interpretability, for in this field it is crucial to gain insight into the why of predictions, to assure that models are actually learning relevant features instead of spurious correlations. To address this, we propose a deep multiscale convolutional architecture trained on the Medical Information Mart for Intensive Care III (MIMIC-III) for mortality prediction, and the use of concepts from coalitional game theory to construct visual explanations aimed to show how important these inputs are deemed by the network. Results show our model attains a ROC AUC of 0.8735 (± 0.0025) which is competitive with the state of the art of Deep Learning mortality models trained on MIMIC-III data, while remaining interpretable. Supporting code can be found at https://github.com/ williamcaicedo/ISeeU.
Keywords: Deep learning | MIMIC-III | ICU | Shapley Values
Complete null agent for games with externalities
عامل پوچ کامل برای بازی با اثرات جانبی-2019
Game theory provides valuable tools to examine expert multi-agent systems. In a cooperative game, col- laboration among agents leads to better outcomes. The most important solution for such games is the Shapley value, that coincides with the expected marginal contribution assuming equiprobability. This as- sumption is not plausible when externalities are present in an expert system. Generalizing the concept of marginal contributions, we propose a new family of Shapley values for situations with externalities. The properties of the Shapley value offer a rationale for its application. This family of values is charac- terized by extensions of Shapley’s axioms: efficiency, additivity, symmetry, and the null player property. The first three axioms have widely accepted generalizations to the framework of games with externali- ties. However, different concepts of null players have been proposed in the literature and we contribute to this debate with a new one. The null player property that we use is weaker than the others. Finally, we present one particular value of the family, new in the literature, and characterize it by two additional properties.
Keywords: Game theory | Multi-agent systems | Externalities | Partition function | Marginal contribution
Hotel location when competitors may react: A game-theoretic gravitational model
مکان هتل وقتی که رقبا ممکن است واکنش نشان دهند: یک مدل نظری گرانشی بازی-2018
This paper presents a hotel location model that incorporates concepts from both game theory and gravitational site location models. We consider a hotel chain intending to build new hotels in a given region. Customers travel to the region to visit some specific points, termed “attractions”, and they choose a hotel according to room price, location and hotel attractiveness. Competitor hotels react to the new hotels by changing prices, in order to maximize their own profits, so the final set of prices will be a Nash equilibrium. We propose an iterative procedure for finding the equilibrium prices and a genetic algorithm-based procedure for finding the optimal strategy, in terms of new hotels to be built and respective typologies. Using a mini case, we illustrate and analyse the influence of several parameters. Then, we present computational experiments, concluding that the proposed procedures are effective in finding good solutions for the model.
keywords: Tourism site location| Game theory| Genetic algorithms| Spatial interaction models
Competitive store closing during an economic downturn
بستن انبار رقابتی درطی یک رکود اقتصادی-2018
The economic downturn of the late 2000s resulted in the closing of a number of retail stores. When doing so a firm has to consider which stores to close, and when to close them, so that profits are maximized. Since retail stores operate in a competitive environment, these decisions are not simply a function of a retailers existing store locations but are also contingent on the location, and closing decisions, of stores operated by rival firms. This paper examines a game between two retail chains looking to downsize operations in a given region and presents a solution procedure that captures the equilibrium store closing decisions. The solution procedure includes a single period mixed integer program model and a multi-period heuristic with a backward and forward pass to find near optimal solutions. Our results provide guidelines for developing effective strategies to systematically reduce the number of stores so that profit is maximized while competitive pressure is exerted on rival stores.
keywords: Location analysis |Store closing |Integer linear programming |Game theory
Psychological contract model for knowledge collaboration in virtual community of practice: An analysis based on the game theory
مدل قرارداد روانشناختی برای همکاری دانش درجامعه مجازی عمل: تجزیه و تحلیل بر اساس بازی تئوری-2018
In virtual communities of practice, many participants use their knowledge to achieve a common goal based on cooperation, and the key to such cooperation is knowledge collab oration. Knowledge collaboration is the primary method of increasing the virtual commu nity of practice’s knowledge ability and achieving the core competency advantage of sus tainable growth. Knowledge collaboration inevitably involves the psychological elements of the collaborators. This study focuses on the elements of the psychological contract of the members of the virtual community of practice when participating in knowledge collabora tion; the psychological contract summarizes the members’ collaboration via seven psycho logical factors spread across two dimensions. Based on these factors, this study uses the game theory to create a cost game model and profit-sharing game model in the virtual community of practice knowledge collaboration. The cost game model is built upon Stack elberg equilibrium, the model is solved using backward induction, and its effectiveness is established. The profit-sharing model is based upon a modified Nash bargaining solution; the model is applied to the Python software development team in the Github community of practice and yields positive results. Finally, the research outcomes are summarized, and directions for future research are provided.
Keywords: Knowledge collaboration ، Virtual community of practice ، Game theory ، Psychological contracts
A game-theoretic approach to optimize the Time-of-Use pricing considering customer behaviors
یک دیدگاه نظریه بازی برای بهینه سازی قیمت گذاری بر اساس زمان استفاده با نظر به رفتارهای مشتری-2018
Time-of-Use (TOU) pricing is an electricity demand response program with great potential to shave the peak demand and eliminate the need for extra power plants. It is necessary to distinguish which TOU program will lead to profitable results to the utility and customers when the customer behaviors and real work shifts are considered. In this paper, a game-theoretic approach is introduced to formulate the problem using a two-layer mathematical programming model, which is solved by the backward induction to gain the Nash-equilibrium. The case study results show that the equilibrium can create a win-win situation for the utility and customers, i.e. the utility increases its profit and the customer reduces its cost. Among different customers, the utility can focus on the customer with a small penalty factor and a large auxiliary coefficient, who prefers to join the TOU program. Even only a small portion of customers joins the TOU program, large improvements in utilitys profit can be made.
keywords: Time-of-Use pricing |Customer behaviors |Game theory |Sustainability
Buffer space hedging and coordination in prefabricated construction supply chain management
مزایده گذاری فضای میانی و هماهنگی در مدیریت زنجیره تامین سازه پیش ساخته-2018
This paper studies a coordination scheme to solve a buffer space hedging (BSH) issue in the prefabricated construction supply chain management (PCSCM). To hedge against unfavorable impacts caused by improper delivery of prefab, the project contractor requires the transportation company to reserve some buffer space in its intermediate warehouse for contingent use (holding safety stock and/or keeping idle). This is termed as BSH strategy. However, this strategy may impact the operation efficiency and potential profit loss for the transportation company. A balance must be kept through a BSH coordination mechanism. Two terms are involved in this mechanism: a BSH amount related cost term is charged by the transportation company to the project contractor and a constant transfer term is used to fairly allocate the system surplus. Three models with different power structures are used to examine the balance. For uneven power settings, two Stackelberg games with alternative decision making sequences are studied. For the equal power setting, a Nash game is presented. It is observed that the proposed coordination mechanism is able to reduce the required BSH amount with win-win coordination. Some interesting managerial implications are also obtained from comparison analysis and numerical studies.
keywords: Prefabricated construction |Supply chain management |Buffer space hedging |Game theory |Supply chain coordination
Conditions of reverse bullwhip effect in pricing under joint decision of replenishment and pricing
شرایط تاثیر شلاقی معکوس در قیمت گذاری تحت تصمیم مشترک تجدید تدارکات و قیمت گذاری-2018
A “reverse bullwhip effect in pricing (RBP)” occurs when an amplification of price variability takes place moving from the upstream suppliers to the downstream customers in a supply chain. In this study, we investigate RBP conditions for supply chains where joint replenishment and pricing decisions are made. Commencing with a single-stage supply chain in which a retailer faces a random and price-sensitive demand, we extend the results to a multi-stage supply chain using a leader-follower game theoretical framework. We discuss RBP conditions for supply chains where newsvendor and continuous review inventory policies are employed, and present numerical examples for commonly used demand functions.
keywords: Pricing |Inventory control |Newsvendor |Continuous review |Bullwhip effect |Supply chain management |Game theory
The economics of attitudes: A different approach to utility functions of players in tourism marketing coalitional networks
اقتصادهای برخوردها: یک دیدگاه متفاوت برای توابع استفاده بازیگران در شبکه های ائتلافی بازاریابی گردشگری-2018
The foundation of destination collaboration is based on the interdependency of the organizations involved in producing destination products. The high rate of destination collaboration failure underscores the need for conflict studies. Unlike previous studies, which depend solely on the collaboration monetary values, this study proposes a new approach to define its utility functions based on the attitudinal and motivational values. We employ the network theory to define the utility function of four major players and the game theory to examine three distribution solutions of coalitional activities values. The results support the notion of “free riders” mentioned in collaboration studies and explains why free riding is a natural phenomenon in tourism destinations’ marketing activities. The findings suggest that individual entities and hospitality are the two players with the highest admission fee and the least contribution. We suggest the concepts of fairness and stability to be considered in incentive policies to encourage collaboration among higher admission players.
keywords: Destination |Collaboration |Free riders |Conflict |Coalitional game |Network theory
Informational and/or transactional websites: Strategic choices in a distribution channel
وب سایت های اطلاعاتی و یا تراکنشی: انتخاب های راهبردی در یک کانال توزیع-2018
While most businesses have faced the decision of whether to operate an informational and/or a transactional website, the literature on website selection in marketing channels remains very sparse. This paper proposes an analytical framework that compares scenarios where a manufacturer uses either an informational, a transactional, or both transactional and informational website in a distribution channel formed by one manufacturer and one retailer. We find that the selection of the optimal website depends on the online market base of the product, the effectiveness of the manufacturer-controlled online communications, and the cross-price effect between online and offline channels. For both the manufacturer and retailer, informational websites are preferable when the online market base is small. With larger online markets, the manufacturer may prefer either informational and transactional websites or exclusively informational websites, while the retailer is always better off with an exclusively informational website. Theoretical and managerial implications of these findings are discussed.
keywords: Marketing strategies |Multichannel commerce |Distribution channels |Game theory