Paper: arXiv 2408.02355

Authors: Mingshu Li, Bhaskarjit Sarmah, Dhruv Desai, Joshua Rosaler, Snigdha Bhagat, Philip Sommer, Dhagash Mehta

Abstract

Due to the dynamic nature of financial markets, maintaining models that produce precise predictions over time is difficult. Often the goal isn’t just point prediction but determining uncertainty. Quantifying uncertainty, especially the aleatoric uncertainty due to the unpredictable nature of market drivers, helps investors understand varying risk levels. Recently, quantile regression forests (QRF) have emerged as a promising solution: Unlike most basic quantile regression methods that need separate models for each quantile, quantile regression forests estimate the entire conditional distribution of the target variable with a single model, while retaining all the salient features of a typical random forest. We introduce a novel approach to compute quantile regressions from random forests that leverages the proximity (i.e., distance metric) learned by the model and infers the conditional distribution of the target variable. We evaluate the proposed methodology using publicly available datasets and then apply it towards the problem of forecasting the average daily volume of corporate bonds. We show that using quantile regression using Random Forest proximities demonstrates superior performance in approximating conditional target distributions and prediction intervals to the original version of QRF. We also demonstrate that the proposed framework is significantly more computationally efficient than traditional approaches to quantile regressions.

Complexity vs Empirical Score

  • Math Complexity: 5.0/10
  • Empirical Rigor: 8.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper introduces a novel algorithmic modification to quantile regression forests, involving non-trivial statistical mechanics and proximity metrics, but grounds it in practical implementation with computational efficiency benchmarks, dataset applications, and performance comparisons on real-world financial data.

Research Flowchart

  flowchart TD
  A["<b>Research Goal</b><br>Quantify uncertainty in corporate bond<br>volume forecasting via novel<br>Random Forest proximities"] --> B["<b>Methodology: Proximity-Based<br>Quantile Regression</b><br>Leverage RF distance metrics to<br>infer conditional distribution<br>(Single model vs. per-quantile)"]
  B --> C["<b>Data Inputs</b><br>Public Datasets &<br>Corporate Bond Volume Data"]
  C --> D["<b>Computational Process</b><br>Train RF → Compute Proximities<br>→ Aggregate Residuals →<br>Estimate Conditional Quantiles"]
  D --> E["<b>Key Findings</b><br>1. Superior distribution approximation<br>2. Better prediction intervals<br>3. Higher computational efficiency<br>vs. traditional QRF"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes a novel quantile regression method that uses random forest proximities to infer conditional distributions for financial forecasting.

Source sentence
We introduce a novel approach to compute quantile regressions from random forests that leverages the proximity (i.e., distance metric) learned by the model and infers the conditional distribution of the target variable.

Assumes

No explicit assumptions found in the text.

Data

  • Boston Housing · mlbench benchmark public
    Source sentence
    Boston Housing 506 14 mlbench
  • Corporate Bonds Data · MarketAxess TRACE market-prices commercial
    December 1, 2023 to April 4, 2024 · U.S. corporate bond securities
    Source sentence
    The data used in this study comprises 675,861 records, covering the period from December 1, 2023, to April 4, 2024.

Costs

Not applicable The paper reports no trading or execution result.

Finds 3

  • RF-GAP consistently outperformed other methods in terms of quantile loss, showing superior capability to capture the underlying geometry and distribution of the data. performance
    Source sentence
    RF-GAP consistently outperformed other methods in terms of quantile loss, demonstrating its superior capability in capturing the underlying geometry and distribution of the data.
  • Quantile regression using random forest proximities yields narrower prediction intervals and higher accuracy compared to the original QRF on both toy and corporate bond datasets. predictive-accuracy
    Source sentence
    Figure 1: Forecasting Results of Quantile Regression using RF-GAP (Dataset: Boston). (a): Fitted values based on conditional median, (b): prediction interval at 95%.
  • Using the quantile loss split criterion gave the best performance at the 50th, 95th, and 97.5th quantiles, though it did not always reduce loss at those quantiles across datasets. statistical
    Source sentence
    quantiles, which in turn yields that a criterion minimizing certain quantile loss does not necessarily lead to reduced quantile loss at that quantile.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 16 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 1 paper in the archive

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.