Source author record

Daniel Fraiman

Daniel Fraiman appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Neurons and Cognition cond-mat.dis-nn Applications Machine Learning math.ST physics.soc-ph q-fin.CP q-fin.ST

Catalog footprint

What is connected

6works

8topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2026arXiv

Imbalanced Classification under Capacity Constraints

In many classification settings, the class of primary interest is underrepresented, leading to imbalanced data problems that arise in applications such as rare disease detection and fraud identification. In these contexts, identifying a potential positive instance typically triggers costly follow-up actions, such as medical imaging or detailed transaction inspection, which are subject to limited operational capacity. Motivated by this setting, we consider classification problems where data may arrive sequentially and decisions must be made under constraints on the number of instances that can be selected for further analysis. We propose a classification framework that explicitly controls the rate of positive predictions, enforcing a user-defined bound on the proportion of observations classified as belonging to the minority class while maximizing detection performance. The approach can be implemented using standard learning methods and naturally extends to online settings, where decisions are taken in real time. We show that incorporating capacity constraints leads to substantial improvements over classical approaches, including resampling techniques such as SMOTE, which do not directly control the selection rate.

preprint2015arXiv

A test of hypotheses for random graph distributions built from EEG data

The theory of random graphs is being applied in recent years to model neural interactions in the brain. While the probabilistic properties of random graphs has been extensively studied in the literature, the development of statistical inference methods for this class of objects has received less attention. In this work we propose a non-parametric test of hypotheses to test if two samples of random graphs were originated from the same probability distribution. We show how to compute efficiently the test statistic and we study its performance on simulated data. We apply the test to compare graphs of brain functional network interactions built from electroencephalographic (EEG) data collected during the visualization of point light displays depicting human locomotion.

preprint2015arXiv

Banking Networks and Leverage Dependence: Evidence from Selected Emerging Countries

We use bank-level balance sheet data from 2005 to 2010 to study interactions within the banking system of five emerging countries: Argentina, Brazil, Mexico, South Africa, and Taiwan. For each country we construct a financial network based on the leverage ratio dependence between each pair of banks, and find results that are comparable across countries. Banks present a variety of leverage ratio behaviors. This leverage diversity produces financial networks that exhibit a modular structure characterized by one large bank community, some small ones and isolated banks. There exist compact structures that have synchronized dynamics. Many groups of banks merge together creating a financial network topology that converges to a unique big cluster at a relatively low leverage dependence level. Finally, we propose a model that includes corporate and interbank loans for studying the banking system. This model generates networks similar to the empirical ones. Moreover, we find that faster-growing banks tend to be more highly interconnected between them, and this is also observed in empirical data.

preprint2012arXiv

What kind of noise is brain noise: anomalous scaling behavior of the resting brain activity fluctuations

The continuous interaction between brain regions "at rest" defines the so-called resting state networks (RSN) which can be reconstructed from the analysis of functional magnetic resonance imaging (fMRI) data. What dynamical mechanism allows for a flexible large-scale organization of the RSN still remains an important challenge. Here, three key novel properties of the RSN are uncovered. First, the correlation length (i.e., the length at which correlation between two regions vanishes) diverges with the cluster's size considered. Second, this divergence it is observed also for measures of mutual information. Third, the variance of the fMRI mean signal remains constant across the entire range of observed clusters sizes, in contrast with naive expectations. The unveiled scale invariance exposes the RSN optimal information-sharing properties across very diverse networks sizes, architectures and functions, which can be an important marker of healthy brain dynamics.

preprint2011arXiv

Point process analysis of large-scale brain fMRI dynamics

Functional magnetic resonance imaging (fMRI) techniques have contributed significantly to our understanding of brain function. Current methods are based on the analysis of \emph{gradual and continuous} changes in the brain blood oxygenated level dependent (BOLD) signal. Departing from that approach, recent work has shown that equivalent results can be obtained by inspecting only the relatively large amplitude BOLD signal peaks, suggesting that relevant information can be condensed in \emph{discrete} events. This idea is further explored here to demonstrate how brain dynamics at resting state can be captured just by the timing and location of such events, i.e., in terms of a spatiotemporal point process. As a proof of principle, we show that the resting state networks (RSN) maps can be extracted from such point processes. Furthermore, the analysis uncovers avalanches of activity which are ruled by the same dynamical and statistical properties described previously for neuronal events at smaller scales. Given the demonstrated functional relevance of the resting state brain dynamics, its representation as a discrete process might facilitate large scale analysis of brain function both in health and disease.

preprint2008arXiv

Ising-like dynamics in large-scale functional brain networks

Brain "rest" is defined -more or less unsuccessfully- as the state in which there is no explicit brain input or output. This work focuss on the question of whether such state can be comparable to any known \emph{dynamical} state. For that purpose, correlation networks from human brain Functional Magnetic Resonance Imaging (fMRI) are constrasted with correlation networks extracted from numerical simulations of the Ising model in 2D, at different temperatures. For the critical temperature $T_c$, striking similarities appear in the most relevant statistical properties, making the two networks indistinguishable from each other. These results are interpreted here as lending support to the conjecture that the dynamics of the functioning brain is near a critical point.