With new treatments and novel technology available, personalized medicine has become a key topic in the new era of healthcare. Traditional statistical methods for personalized medicine and subgroup identification primarily focus on single treatment or two arm randomized control trials (RCTs). With restricted inclusion and exclusion criteria, data from RCTs may not reflect real world treatment effectiveness. However, electronic medical records (EMR) offers an alternative venue. In this paper, we propose a general framework to identify individualized treatment rule (ITR), which connects the subgroup identification methods and ITR. It is applicable to both RCT and EMR data. Given the large scale of EMR datasets, we develop a recursive partitioning algorithm to solve the problem (ITR-Tree). A variable importance measure is also developed for personalized medicine using random forest. We demonstrate our method through simulations, and apply ITR-Tree to datasets from diabetes studies using both RCT and EMR data. Software package is available at https://github.com/jinjinzhou/ITR.Tree.

Thesis deals with Prospect Theory and with its findings when making decisions under risk which is called four-fold pattern. In theoretical part is shortly mentioned the field of Behavioral finance and further described original Prospect Theory and also its modified version called Cumulative Prospect Theory. The main goal of the practical part is to confirm on the basis of questionnaire between different kinds of college graduates the four-fold pattern, identify between 3 segments differences and make the final conclusion whether the type of college education has an impact on four-fold pattern.

<p>The field of reinforcement learning concerns the question of automated action selection given past experiences. As an agent moves through the state space, it must recognize which state choices are best in terms of allowing it to reach its goal. This is quantified with value functions, which evaluate a state and return the sum of rewards the agent can expect to receive from that state. Given a good value function, the agent can choose the actions which maximize this sum of rewards. Value functions are often chosen from a linear space defined by a set of features; this method offers a concise structure, low computational effort, and resistance to overfitting. However, because the number of features is small, this method depends heavily on these few features being expressive and useful, making the selection of these features a core problem. This document discusses this selection.</p><p>Aside from a review of the field, contributions include a new understanding of the role approximate models play in value function approximation, leading to new methods for analyzing feature sets in an intuitive way, both using the linear and the related kernelized approximation architectures. Additionally, we present a new method for automatically choosing features during value function approximation which has a bounded approximation error and produces superior policies, even in extremely noisy domains.</p> / Dissertation

No description available.

Kahneman and Tversky (1979) first demonstrated that when individuals decide whether or not to accept a gamble, potential losses receive more weight than possible gains in the decision. This phenomenon is referred to as loss aversion. We investigated how loss aversion in risky financial decisions is influenced by sudden changes to wealth, employing both behavioral and neurobiological measures. We implemented an fMRI experimental paradigm, based on that employed by Tom et al. (2007). There are two treatments, called RANDOM and CONTINGENT. In RANDOM, the baseline setting, the changes to wealth, referred to as wealth shocks in economics, are independent of the actual choices participants make. Under CONTINGENT, we induce the belief that the changes in income are a consequence of subjects' own decisions. The magnitudes and sequence of the shocks to wealth are identical between the CONTINGENT and RANDOM treatments. We investigated whether more loss aversion existed in one treatment than another. The behavioral results showed significantly greater loss aversion in CONTINGENT compared to RANDOM after a negative wealth shock. No differences were observed in the response to positive shocks. The fMRI results revealed a neural loss aversion network, comprising the bilateral striatum, amygdala and dorsal anterior cingulate cortex that was common to the CONTINGENT and RANDOM tasks. However, the ventral prefrontal cortex, primary somatosensory cortex and superior occipital cortex, showed greater activation in response to a negative change in wealth due to individual's own decisions than when the change was exogenous. These results indicate that striatum activation correlates with loss aversion independently of the source of the shock, and that the ventral prefrontal cortex (vPFC) codes the experimental manipulation of agency in one's actions influencing loss aversion.

No description available.

We consider a single-product dynamic inventory problem where the demand distributions in each period are known and independent but with density. We assume the lead time and the fixed cost for ordering are zero and that there are no capacity constraints. There is a holding cost and a backorder cost for unfulfilled demand, which is backlogged until it is filled by another order. The problem may be nonstationary, and in fact our approximation of the optimal cost function using splines is most advantageous when demand falls suddenly. In this case the myopic policy, which is most often used in practice to calculate optimal inventory level, would be very costly. Our algorithm uses quadratic splines to approximate the optimal cost function for this dynamic inventory problem and calculates the optimal inventory level and optimal cost.

The paper describes the software implementation of the numerical method proposed by M.Bardi and M.Falcone solving optimal time games. Examples of numerical calculations are given. The question of the applicability of this method for solving differential games with life line, i.e. with a set where the second player escapes and wins unconditionally, is discussed and examined. Currently, the study has not been completed and will be continued in the future. / В работе приведено описание программной реализации численного метода, предложенного М.Барди и М.Фальконе для решения игр быстродействия. Приведены примеры численного счета. Обсуждается и исследуется вопрос о применимости данного метода для решения дифференциальных игр быстродействия с линией жизни, то есть с множеством, при попадании системы на которое второй игрок безусловно выигрывает. В настоящее время это исследование не доведено до конца и будет продолжено в дальнейшем.

Made available in DSpace on 2010-04-20T21:00:18Z (GMT). No. of bitstreams: 3 martincasalsturma2003.pdf.jpg: 11517 bytes, checksum: 930c75c74ff8269d877983110854646d (MD5) martincasalsturma2003.pdf: 975293 bytes, checksum: f41cfd3a5f3e659d07e2861acf4e23d4 (MD5) martincasalsturma2003.pdf.txt: 107784 bytes, checksum: e96ad0fb809a12f13ba7d3f08e24c1ee (MD5) Previous issue date: 2006-02-15T00:00:00Z / O objetivo deste trabalho é analisar a alocação de investimentos no mercado acionário brasileiro, utilizando a teoria do prospecto de Tversky e Kahneman (1979) e o conceito de Aversão a Perdas Míope (Myopic Loss Aversion) proposto por Benartzi e Thaler (1995). Foram levantados através de experimento de laboratório os parâmetros da função de valor e da função de ponderação de probabilidades da teoria do prospecto e foi verificada a alocação de investimentos entre ações e renda fixa que maximizam a utilidade. Chegamos à conclusão que o total de recursos atualmente direcionados ao mercado de ações no Brasil, que é de aproximadamente 2,7% para pessoas físicas e de 6,0% para pessoas jurídicas, é compatível com a teoria do prospecto. / The objective of this study is to analyze the investment allocation in the Brazilian stock market, using Tversky and Kahneman’s prospect theory (1979) and the concept of myopic loss aversion proposed by Benartzi and Thaler (1995). We run a laboratory experiment to obtain the parameters of the value function and the probability weighting function of the prospect theory and identify the allocation that maximizes utility in the Brazilian Market We conclude that the actual allocation of investment in the stock market, of around 2.7% for individuals and around 6% for all the segments, is in accordance with the prospect theory.