Data Science (DS)
http://hdl.handle.net/20.500.11824/9
2021-09-08T12:09:53ZOn solving cycle problems with Branch-and-Cut: extending shrinking and exact subcycle elimination separation algorithms
http://hdl.handle.net/20.500.11824/1326
On solving cycle problems with Branch-and-Cut: extending shrinking and exact subcycle elimination separation algorithms
Kobeaga, G.; Merino, M.; Lozano, J.A.
In this paper, we extend techniques developed in the context of the Travelling Salesperson Problem for cycle problems. Particularly, we study the shrinking of support graphs and the exact algorithms for subcycle elimination separation problems. The efficient application of the considered techniques has proved to be essential in the Travelling Salesperson Problem when solving large size problems by Branch-and-Cut, and this has been the motivation behind this work. Regarding the shrinking of support graphs, we prove the validity of the Padberg–Rinaldi general shrinking rules and the Crowder–Padberg subcycle-safe shrinking rules. Concerning the subcycle separation problems, we extend two exact separation algorithms, the Dynamic Hong and the Extended Padberg–Grötschel algorithms, which are shown to be superior to the ones used so far in the literature of cycle problems. The proposed techniques are empirically tested in 24 subcycle elimination problem instances generated by solving the Orienteering Problem (involving up to 15,112 vertices) with Branch-and-Cut. The experiments suggest the relevance of the proposed techniques for cycle problems. The obtained average speedup for the subcycle separation problems in the Orienteering Problem when the proposed techniques are used together is around 50 times in medium-sized instances and around 250 times in large-sized instances.
2021-01-01T00:00:00ZStatistical assessment of experimental results: a graphical approach for comparing algorithms
http://hdl.handle.net/20.500.11824/1323
Statistical assessment of experimental results: a graphical approach for comparing algorithms
Arza, E.; Ceberio, J.; Irurozki, E.; Pérez, A.
Non-deterministic measurements are common in real-world scenarios: the performance of a stochastic optimization algorithm or the total reward of a reinforcement learning agent in a chaotic environment are just two examples in which unpredictable outcomes are common. These measures can be modeled as random variables and compared among each other via their expected values or more sophisticated tools such as null hypothesis statistical tests. In this paper, we propose an alternative framework to compare two random variables according to their cumulative distribution functions. First, we introduce a dominance measure for two random variables that quantifies the proportion in which the cumulative distribution function of one of the random variables is greater than the other. Then, we present a graphical method that allows a visual estimation of the proposed dominance measure, the probability that one of the random variables takes lower values than the other, and a comparison of quantiles of the random variables. With illustrative purposes, we re-evaluate the experimentation of an already published work with the proposed methodology and we show that additional conclusions—missed by the rest of the methods—can be inferred. Additionally, a software package is provided as a convenient way of applying the proposed framework.
2021-08-25T00:00:00ZSimulation approach for assessing the performance of the γEWMA control chart
http://hdl.handle.net/20.500.11824/1320
Simulation approach for assessing the performance of the γEWMA control chart
Patino-Rodriguez, C.; Pérez, D. M.; Usuga Manco, O.
i) Purpose: The purpose of this paper is to evaluate the performance of a modified EWMA control chart ($\gamma$EWMA control chart), which considers data distribution and incorporate its correlation structure, simulating in-control and out-of-control processes and to select an adequate value for smoothing parameter with these conditions. ii) Design/methodology/approach:
This paper is based on a simulation approach using the methodology for evaluating statistical methods proposed by Morris et al. (2019). Data were generated from a simulation considering two factors that associated with data: (1) quality variable distribution skewness as an indicator of quality variable distribution; (2) the autocorrelation structure for type of relationship between the observations and modeled by AR(1). In addition, one factor associated with the process was considered, (1) the shift in the process mean. In the following step, when the chart control is modeled, the fourth factor intervenes. This factor is a smoothing parameter. Finally, three indicators defined from the Run Length are used to evaluate γEWMA control chart performance this factors and their interactions. iii) Findings: Interaction analysis for four factor evidence that the modeling and selection of parameters is different for out-of-control and in-control processes therefore the considerations and parameters selected for each case must be carefully analyzed. For out-of-control processes, it is better to preserve the original features of the distribution (mean and variance) for the calculation of the control limits. It makes sense that highly autocorrelated observations require smaller smoothing parameter since the correlation structure enables the preservation of relevant information in past data. iv) Originality/value: The $\gamma$EWMA control chart there has advantages because it gathers, in single chart control: the process and modelling characteristics, and data structure process. Although there are other proposals for modified EWMA, none of them simultaneously analyze the four factors nor their interactions. The proposed $\gamma$EWMA allows setting the appropriate smoothing parameter when these three factors are considered.
2021-02-22T00:00:00ZAltered effective connectivity in sensorimotor cortices: a novel signature of severity and clinical course in depression
http://hdl.handle.net/20.500.11824/1313
Altered effective connectivity in sensorimotor cortices: a novel signature of severity and clinical course in depression
Ray, D.; Bezmaternykh, D.; Mel'nikov, M.; Friston, K.J.; Das, M.
Functional neuroimaging research on depression has traditionally targeted neural networks associated with the psychological aspects of depression. In this study, instead, we focus on alterations of sensorimotor function in depression. We used resting-state functional MRI data and Dynamic Causal Modeling (DCM) to assess the hypothesis that depression is associated with aberrant effective connectivity within and between key regions in the sensorimotor hierarchy. Using hierarchical modeling of between-subject effects in DCM with Parametric Empirical Bayes we first established the architecture of effective connectivity in sensorimotor cortices. We found that in (interoceptive and exteroceptive) sensory cortices across participants, the backward connections are predominantly inhibitory whereas the forward connections are mainly excitatory in nature. In motor cortices these parities were reversed. With increasing depression severity, these patterns are depreciated in exteroceptive and motor cortices and augmented in the interoceptive cortex: an observation that speaks to depressive symptomatology. We established the robustness of these results in a leave-one-out cross validation analysis and by reproducing the main results in a follow-up dataset. Interestingly, with (non-pharmacological) treatment, depression associated changes in backward and forward effective connectivity partially reverted to group mean levels. Overall, altered effective connectivity in sensorimotor cortices emerges as a promising and quantifiable candidate marker of depression severity and treatment response.
2021-01-01T00:00:00Z