Single Decision Rule Research Articles

Providing Evidence for the Null Hypothesis in Functional Magnetic Resonance Imaging Using Group-Level Bayesian Inference.

Classical null hypothesis significance testing is limited to the rejection of the point-null hypothesis; it does not allow the interpretation of non-significant results. This leads to a bias against the null hypothesis. Herein, we discuss statistical approaches to ‘null effect’ assessment focusing on the Bayesian parameter inference (BPI). Although Bayesian methods have been theoretically elaborated and implemented in common neuroimaging software packages, they are not widely used for ‘null effect’ assessment. BPI considers the posterior probability of finding the effect within or outside the region of practical equivalence to the null value. It can be used to find both ‘activated/deactivated’ and ‘not activated’ voxels or to indicate that the obtained data are not sufficient using a single decision rule. It also allows to evaluate the data as the sample size increases and decide to stop the experiment if the obtained data are sufficient to make a confident inference. To demonstrate the advantages of using BPI for fMRI data group analysis, we compare it with classical null hypothesis significance testing on empirical data. We also use simulated data to show how BPI performs under different effect sizes, noise levels, noise distributions and sample sizes. Finally, we consider the problem of defining the region of practical equivalence for BPI and discuss possible applications of BPI in fMRI studies. To facilitate ‘null effect’ assessment for fMRI practitioners, we provide Statistical Parametric Mapping 12 based toolbox for Bayesian inference.

Outcome-Weighted Learning for Personalized Medicine with Multiple Treatment Options.

To achieve personalized medicine, an individualized treatment strategy assigning treatment based on an individual's characteristics that leads to the largest benefit can be considered. Recently, a machine learning approach, O-learning, has been proposed to estimate an optimal individualized treatment rule (ITR), but it is developed to make binary decisions and thus limited to compare two treatments. When many treatment options are available, existing methods need to be adapted by transforming a multiple treatment selection problem into multiple binary treatment selections, for example, via one-vs-one or one-vs-all comparisons. However, combining multiple binary treatment selection rules into a single decision rule requires careful consideration, because it is known in the multicategory learning literature that some approaches may lead to ambiguous decision rules. In this work, we propose a novel and efficient method to generalize outcome-weighted learning for binary treatment to multi-treatment settings. We solve a multiple treatment selection problem via sequential weighted support vector machines. We prove that the resulting ITR is Fisher consistent and obtain the convergence rate of the estimated value function to the true optimal value, i.e., the estimated treatment rule leads to the maximal benefit when the data size goes to infinity. We conduct simulations to demonstrate that the proposed method has superior performance in terms of lower mis-allocation rates and improved expected values. An application to a three-arm randomized trial of major depressive disorder shows that an ITR tailored to individual patient's expectancy of treatment efficacy, their baseline depression severity and other characteristics reduces depressive symptoms more than non-personalized treatment strategies (e.g., treating all patients with combined pharmacotherapy and psychotherapy).

LEMRG: Decision Rule Generation Algorithm for Mining MicroRNA Expression Data.

Recently, research on mining microRNA (or miRNA) expression data has received a lot of attention, mainly because of its role in gene regulation. However, such type of data - usually saved in the form of microarrays - are very specific, because they contain only a small number of cases (often less than 100) compared with large number of attributes (equal to several hundreds or even tens of thousand). The small number of cases available during the learning process can cause instability of the newly created classifiers. Secondly, the huge number of attributes imposes the necessity of selecting only a few dominant attributes strongly correlated with the decision. Thus, an application of fundamental machine learning approaches of mining microarray data and its further classification is problematic or even could just fail.Thus, the main goal of our research is to develop the generalized algorithm of mining microarray data (including miRNA data sets), mainly to improve stability and, consequently, accuracy of classification for the newly created learning classifiers. The main concept of the novel approach is based on iteratively inducing many subsequent decision rule sets - called decision rule generations - instead of inducing only a single decision rule set, as it is done routinely. The decision rules have been chosen as the baseline classifiers of the newly developed LEMRG (Learning from Examples Module based on Rule Generations) algorithm mainly because the decision rule-based knowledge representation is easier for humans to comprehend, rather than other learning models. In our research we used a miRNA expression level learning data set describing 11 types of human cancers, while the testing data set contained poorly differentiated cases of only four types of cancers. As expected, our new classifiers - saved in the form of so-called cumulative decision rule sets - had better stability and accuracy of classification than single decision rule sets induced in the traditional manner. Furthermore, the LEMRG was compared with other machine learning models. It was proven that only 3 out of all 16 tested classifiers enabled so effective classification as our newly developed approach. Thus, using our cumulative set of decision rules, all cases of cancer from two selected concepts - colon and ovary - were correctly classified. Furthermore, we showed the role of these selected miRNAs as the potential biomarkers for diagnosis of tumors.A preliminary result of our research on decision rule generations was initially presented at the first International Conference of Digital Medicine and Medical 3D Printing (17-19.06.2016, Nanjing, China).

Routing Diverse Evacuees with the Cognitive Packet Network Algorithm

Regarding mobility, health conditions and personal preferences, evacuees can be categorized into different classes in realistic environments. Previous emergency navigation algorithms that direct evacuees with a single decision rule cannot fulfil civilians’ distinct service requirements and increase the likelihood of inducing destructive crowd behaviours, such as clogging, pushing and trampling, due to diverse mobility. This paper explores a distributed emergency navigation algorithm that employs the cognitive packet network concept to tailor different quality of service needs to different categories of evacuees. In addition, a congestion-aware algorithm is presented to predict the future congestion degree of a path with respect to the observed population density, arrival rate and service rate of each route segment. Experiments are implemented in a simulated environment populated with autonomous agents. Results show that our algorithm can increase the number of survivors while providing improved quality of service.

A parsimonious optimal foraging model explaining mortality patterns in Serengeti wildebeest

Based on data collected over 24 years in the Serengeti in Tanzania, Sinclair and Arcese (1995) indicated that the sensitivity of blue wildebeest Connochaetes taurinus to predation risk by lions Panthera leo may cause them to change habitats between open (low risk) and wooded (risky) habitats. They found that, in poor rainfall years, predators kill wildebeest that are in better condition than those that die of natural causes. In good rainfall years, predators kill wildebeest that are in worse condition than those that die of natural causes. Sinclair and Arcese (1995) proposed the “predation-sensitive food” hypothesis. This hypothesis suggests that, as food becomes limiting, animals take greater risks to obtain more food, and some of these animals are killed. I propose a more parsimonious hypothesis based on the marginal value theorem that is consistent with the observations made by Sinclair and Arcese (1995). Wildebeest follow a single decision rule in good and poor rainfall years, viz. move when foraging elsewhere increases your rate of intake of nutritious food. Similarly, predators follow a single decision rule in good and poor rainfall years, viz. take the prey item that maximizes the intake of energy per unit effort expended. This parsimonious model does not require differences in predator sensitivity as required by Sinclair and Arcese's (1995) model. I indicate ways in which my model can be falsified.

Economics and euthanasia

This paper argues that a cost/benefit-type economic analysis of whether or not to prohibit euthanasia should not be undertaken by health services managers unless ethical issues have been resolved to a particular conclusion. A conclusion on ethical grounds in favour of the prohibition of euthanasia renders the consideration of economic criteria redundant. The case of the German euthanasia programme in the 1930s is used to illustrate the danger of attempting to combine ethical and economic criteria into a single decision rule.

High-Dose Chemotherapy in the Adjuvant Treatment of Breast Cancer: Benefit/Risk Analysis.

BACKGROUND: High-dose chemotherapy (HDRx) may improve the prognosis of patients with high-risk breast cancer but at the expense of increased toxicity. However, no randomized, controlled trials have been published that clearly demonstrate the superiority of HDRx over conventional adjuvant chemotherapy. METHODS: We developed a simple model to compare benefits and risks of HDRx with conventional adjuvant chemotherapy (SDRx). The model integrates data on efficacy and risks of two competing treatment strategies into a single decision rule. RESULTS AND CONCLUSIONS: Using data from phase II studies, we show that if a disease-free survival is considered to be the most important outcome, HDRx should be administered when the probability of breast cancer relapse at five years exceeds 54% to 71% for patients with 4 to 9 positive nodes and exceeds 29% to 40% for patients with >9 positive nodes. If the endpoint of interest is five-year overall survival, then the treatment should be administered when the probability of relapse exceeds 77% to 83% for patients with 4 to 9 nodes involved and 22% to 31% for those with >9 lymph nodes involved. While awaiting results of randomized, controlled trials to definitively establish the efficacy rate of HDRx, we also found that HDRx could be considered in the management of high-risk breast cancer if its efficacy rate is at least 54% to 60% superior to SDRx in reducing relapse risk in breast cancer patients with 4 to 9 nodes and at least 31% to 38% for >9 positive nodes. If survival data are used instead of disease-free survival outcomes, HDRx efficacy should be at least 47% to 48% superior to SDRx in reducing death risk in breast cancer patients with 4 to 9 nodes and at least 27% to 30% superior for >9 positive nodes to consider its use in the adjuvant setting.

Interaction effects on MRP nervousness

SUMMARY Relative performance of decision rules for specific production and inventory control problems has gained attention in the last two decades. One of these areas receiving attention is MRP nervousness. Generally, the issue of MRP nervousness has been examined under highly simplified environments for single decision rule effects. Implications of the results of these previous studies may be misleading in light of interaction effects with different decision rules. Conclusions drawn may not be appropriate when interactions of decision rules are considered under more realistic environments. Statistical interactions have been found and reported (Minifie and Heard 19S5). This paper addresses the managerial implications of these statistical findings and highlights the importance of various features that should be present in simulated MRP environments.

An Outline for a Model of Party Choice

This article is a synthetic effort. It attempts to mold the issue voter of traditional democratic theory and choice models with the nature of the times and partisan voters of empirical voting studies. The vehicle for this attempted synthesis is a voter decision rule more complex, more inclusive, and perhaps less rational than others previously suggested. After motivating and developing the formal representation of the decision rule, a variety of empirical findings about voting behavior are reexamined in light of the rule. Some reinterpretations result, particularly in the case of party identification. This article takes a first step toward formalizing a number of ideas about American voting behavior. Specifically, I will sketch a model which integrates the familiar notions of party identification, retrospective voting, and issue voting into a single decision rule. As intimated, the ideas in this paper are not new ones. But the process of formalizing familiar notions may force a greater degree of precision in our thinking, and perhaps make evident the full implications of the ways we think about voting behavior (see Fiorina, 1975). The article divides into two sections. In the first I outline a model of the individual voting decision, proceeding from the simplest conception of electoral choice to the most complex. This section is largely abstract. In the second section of the paper I attempt to show that the model provides an explanation for a variety of empirical findings which emerge from the research of specialists in voting behavior. Such findings include low correlations between specific issue positions and the ultimate vote decision, the general importance of party ID for voting behavior and why it may vary over time, the strengthening of party ID with political experience in some societies but not in others, incumbency advantages, the importance of the personal qualities of candidates for office, and others.

Elimination by aspects as a decision rule for risky choice

The Elimination By Aspects (EBA) model of Restle (1961) and Tversky (1972a, b) was applied to risky choice. The choice alternatives were duplex gambles and it was assumed that their valued aspects are explicit probabilities and payoffs. Four consequences of the model were described and three of these were tested in a paired comparisons experiment. The experiment was ran as a computer game, where the computer gave feedback of the results of the gambles chosen. The choices were not hypothetical, since the results of bets determined the amount subjects were paid. Substantial violations of two of the consequences were found while the third was not violated. It was suggested that the violations were due to interactions between the explicit aspects which produced new aspects. It is concluded that (i) the EBA model with explicit aspects cannot hold unless the aspects are independent and (ii) no single decision rule, such as Elimination By Aspects, can account for choice behaviour.