arXiv subjects

Juan Camilo Orduz

Publications and source records attributed to Juan Camilo Orduz.

5 recordsLinked to original sources

arXiv2025

Cohort Revenue & Retention Analysis: A Bayesian Approach

We present a Bayesian approach to model cohort-level retention rates and revenue over time. We use Bayesian additive regression trees (BART) to model the retention component which we couple with a linear model for the revenue component. This method is flexible enough to allow adding additional covariates to both model components. This Bayesian framework allows us to quantify uncertainty in the estimation, understand the effect of covariates on retention through partial dependence plots (PDP) and individual conditional expectation (ICE) plots, and most importantly, forecast future revenue and retention rates with well-calibrated uncertainty through highest density intervals. We also provide alternative approaches to model the retention component using neural networks and inference through stochastic variational inference.

stat.AP↗

arXiv2024

BlackJAX: Composable Bayesian inference in JAX

BlackJAX is a library implementing sampling and variational inference algorithms commonly used in Bayesian computation. It is designed for ease of use, speed, and modularity by taking a functional approach to the algorithms' implementation. BlackJAX is written in Python, using JAX to compile and run NumpPy-like samplers and variational methods on CPUs, GPUs, and TPUs. The library integrates well with probabilistic programming languages by working directly with the (un-normalized) target log density function. BlackJAX is intended as a collection of low-level, composable implementations of basic statistical 'atoms' that can be combined to perform well-defined Bayesian inference, but also provides high-level routines for ease of use. It is designed for users who need cutting-edge methods, researchers who want to create complex sampling methods, and people who want to learn how these work.

cs.MS↗

arXiv2021

Modelling stochastic time delay for regression analysis

Systems with stochastic time delay between the input and output present a number of unique challenges. Time domain noise leads to irregular alignments, obfuscates relationships and attenuates inferred coefficients. To handle these challenges, we introduce a maximum likelihood regression model that regards stochastic time delay as an "error" in the time domain. For a certain subset of problems, by modelling both prediction and time errors it is possible to outperform traditional models. Through a simulated experiment of a univariate problem, we demonstrate results that significantly improve upon Ordinary Least Squares (OLS) regression.

stat.ME↗

arXiv2018

The $S^1$-Equivariant signature for semi-free actions as an index formula

John Lott defined an integer-valued signature $σ_{S^1}(M)$ for the orbit space of a compact orientable manifold with a semi-free $S^1$-action but he did not construct a Dirac-type operator which has this signature as its index. We construct such operator on the orbit space and we show that it is essentially unique and that its index coincides with Lott's signature, at least when the stratified space satisfies the so-called Witt condition. For the non-Witt case, this operator remains essentially self-adjoint (in contrast to the Hodge de-Rham operator) and it has a well defined index which we conjecture will also compute $σ_{S^1}(M)$. This article is a condensed version of the original author's PhD Thesis where the theory of induced Dirac-Schrödinger-type operators is developed in detail.

math.DG↗

arXiv2018

Induced Dirac-Schrödinger operators on $S^1$-semi-free quotients

John Lott has computed an integer-valued signature for the orbit space of a compact orientable $(4k+1)$ manifold with a semi-free $S^1$-action, which is a homotopy invariant of that space, but he did not construct a Dirac type operator which has this signature as its index. In this Thesis, we construct such operator on the orbit space, a Thom-Mather stratified space with one singular stratum of positive dimension, and we show that it is essentially unique and that its index coincides with Lott's signature, at least when the stratified space satisfies the so called Witt condition. We call this operator the induced Dirac-Schrödinger operator. The strategy of the construction is to "push down" an appropriate $S^1$-invariant first order transversally elliptic operator to the quotient space. The Witt condition, a topological condition which in this case depends on the codimension of the fixed point set, has various analytic consequences. In particular, when not satisfied, the Hodge-de Rham operator on the quotient space does not need to be essentially self-adjoint and therefore a choice of boundary conditions is required. This choice freedom is not natural in view of the fact that Lott's signature is well defined independently of the Witt condition. The Dirac-Schrödinger operator constructed in this Thesis differs from the Hodge-de Rham operator by a zero order term which ensures it to be essentially self-adjoint. Moreover, this zero order term anti-commutes with the chirality involution allowing the whole operator to split so that the index can be computed even if the Witt condition is not satisfied.

math.DG↗