Mirsad Hadzikadic

Application of complex adaptive systems to pricing of reproducible information goods

Piracy of copyrighted information goods such as computer software, music recordings, and movies has received increased attention in the literature. Much of this research relied on mathematical modeling to analyze pricing policies, protection against piracy, and government policies. We use complex adaptive systems as an alternative methodology to analyze pricing decisions in an industry with products which can be pirated. This approach has been previously applied to pricing and can capture some aspects of the problem which are difficult to analyze using traditional mathematical modeling. The results indicate that advances in technology make a skimming strategy the least preferable approach for producers. Further, improvements in technology, more specifically data communications and the Internet, will erode the profitability of a skimming strategy. The analysis also indicates that complex adaptive systems may provide a useful method for analyzing problems in which interactions between participants in the systems, i.e. consumers, sellers, and regulating agencies, are important in determining the behavior of the system.

Systems modeling and simulation applications for critical care medicine

Critical care delivery is a complex, expensive, error prone, medical specialty and remains the focal point of major improvement efforts in healthcare delivery. Various modeling and simulation techniques offer unique opportunities to better understand the interactions between clinical physiology and care delivery. The novel insights gained from the systems perspective can then be used to develop and test new treatment strategies and make critical care delivery more efficient and effective. However, modeling and simulation applications in critical care remain underutilized. This article provides an overview of major computer-based simulation techniques as applied to critical care medicine. We provide three application examples of different simulation techniques, including a) pathophysiological model of acute lung injury, b) process modeling of critical care delivery, and c) an agent-based model to study interaction between pathophysiology and healthcare delivery. Finally, we identify certain challenges to, and opportunities for, future research in the area.

An Agent-Based Model of Solid Tumor Progression

Simulation techniques used to generate complex biological models are recognized as promising research tools especially in oncology. Here, we present a computer simulation model that uses an agent-based system to mimic the development and progression of solid tumors. The model includes influences of the tumors own features, the host immune response and level of tumor vascularization. The interactions among those complex systems were modeled using a multi-agent modeling environment provided by Netlogo. The model consists of a hierarchy of active objects including cancer cells, immune cells, and energy availability. The simulations conducted indicate the key importance of the nutrient needs of the tumor cells and of the initial responsiveness of the immune system in the tumor progression. Furthermore, the model strongly suggests that immunotherapy treatment will be efficient in individual with sustained immune responsiveness.

Learning to predict: INC2.5

Discusses INC2.5, an incremental concept formation system. The goal of INC2.5 is to form a hierarchy of concept descriptions based on previously-seen instances which are to be used to predict the classification of a new instance description. Each subtree of the hierarchy consists of instances which are similar to each other. The further from the root, the greater the similarity is between the instances within the same groupings. The ability to classify instances based on their description has many applications. For example, in the medical field, doctors are required daily to diagnose patients, in other words to classify patients according to their symptoms. INC2.5 has been successfully applied to several domains, including breast cancer, general trauma, congressional voting records and the monks problems.

An agent based modeling approach for determining optimal price-rebate schemes

Abstract Delayed incentives in the form of cash mail-in rebates have become very popular. While some research has been conducted on consumer perception and behavior toward rebates, little research has been undertaken with respect to a seller’s optimal rebate strategy. We use an agent-based modeling approach for jointly determining optimal price and rebate value. This approach has been previously applied to pricing and can capture some aspects of the problem which are difficult to analyze using traditional mathematical modeling. The model indicates that rebates are much more profitable when their values are jointly determined with price. In addition, the profitability of rebates at a given time depends on past consumers’ experience with rebates and that consumers’ memory of their past rebate redemption behavior plays an important role in determining the profitability of rebates. Also, the analysis shows that agent-based modeling may provide an alternative method for analyzing problems in which interactions between participants in the systems, i.e. consumers, sellers, and regulating agencies, are important in determining the overall behavior of the system.

Automated design of diagnostic systems

This research effort represents an inquiry into an important problem of automated acquisition, indexing, retrieval, and effective use of knowledge in diagnostic tasks. Its specific goal is to develop an incremental concept formation system which will automate both the design and use of diagnostic knowledge-based systems by a novice. The adopted approach to this problem is based on the modified family resemblance and contrast model theories, as well as a context-sensitive, distributed probabilistic representation of learned concepts. These ideas have been implemented in the INC2 system. The system is evaluated in terms of its prediction accuracy in the domains of breast cancer, primary tumor, and audiology cases.

A Computer Simulation Laboratory for Social Theories

We present an agent-based model that allows the user to employ different social theories to try to explain and predict social changes. The model is set in the context of an armed insurgency in a developing country. We demonstrate the capabilities of the model by showing how it simulates a news report-based scenario under different theories and combinations of theories.

Alternative Formulas for Rating Prediction Using Collaborative Filtering

This paper proposes and evaluates several alternate design choices for common prediction metrics employed by neighborhood-based collaborative filtering approach. It first explores the role of different baseline user averages as the foundation of similarity weighting and rating normalization in prediction, evaluating the results in comparison to traditional neighborhood-based metrics using the MovieLens data set. The approach is further evaluated on the Netflix movie data set, using a baseline correlation formula between movies, without meta-knowledge. For the Netflix domain, the approach is augmented with a significance weighting variant that results in an improvement over the original metric. The resulting approach is shown to improve accuracy for neighborhood-based collaborative filtering, and it is general and applicable to establishing relationships among agents with a common list of items which establish their preferences.

30 Day hospital readmission analysis

Readmissions to a hospital after procedures are costly and considered to be an indication of poor quality. As Per the Affordable Care Act of 2010, hospitals may be reimbursed at a reduced rate for patients readmitted to a hospital within 30 days of discharge. In this project, we used statistical and machine-learning methods to analyze the Nationwide Inpatient Sample dataset provided by HCUP (Healthcare Cost and Utilization Project) to identify various clinical, demographic and socio-economic factors that play crucial roles in predicting the revenue loss due to readmissions. Three medical conditions, namely chronic obstructive pulmonary disorder (COPD), total hip arthroplasty (THA), and total knee arthroplasty (TKA) have been primarily used for this purpose. Our analysis builds on both non-parametric and parametric statistical models and machine learning techniques such as Decision Tree, Gradient Boosting, Logistic Regression and Neural Networks. We evaluated and compared these models based on Area under ROC (AUC) and misclassification rate. By including visual analytics, this analysis not only enables the hospitals to compute the loss of revenue but also monitors their quality of service in a real-time fashion.

Managing Complexity: Practical Considerations in the Development and Application of ABMs to Contemporary Policy Challenges

This book emerged out of a project initiated and funded by the Defense Advanced Research Projects Agency (DARPA) that sought to build on efforts to transform agent-based models into platforms for predicting and evaluating policy responses to real world challenges around the world. It began with the observation that social science theories of human behavior are often used to estimate the consequences of alternative policy responses to important issues and challenges. However, alternative theories that remain subject to contradictory claims are ill suited to inform policy. The vision behind the DARPA project was to mine the social sciences literature for alternative theories of human behavior, and then formalize, instantiate, and integrate them within the context of an agent-based modeling system. The research team developed an experimental platform to evaluate the conditions under which alternative theories and groups of theories applied. The end result was a proof of concept developed from the ground up of social knowledge that could be used as an informative guide for policy analysis. This book describes in detail the process of designing and implementing a pilot system that helped DARPA assess the feasibility of a computational social science project on a large scale.