Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

An Explainable Market Integrity Monitoring System with Multi-Source Attention Signals and Transparent Scoring

Sandeep Neela

TL;DR

AIMM-X tackles market integrity surveillance by enabling auditable, explainable detection using publicly available OHLCV data and multi-source attention proxies. It constructs an interpretable Integrity Score $M$ decomposed into six phi-signals, enabling transparent attribution of why a window is flagged. The study demonstrates feasibility across 24 tickers with 233 detected windows in 2024, highlighting the framework’s potential as a triage tool for compliance, research, and regulatory engagement. While not a definitive detector of manipulation, AIMM-X provides a reproducible, open framework that surface anomalies for expert review, outlines practical deployment considerations, and maps a clear path for data enhancements and validation with ground-truth cases.

Abstract

Market integrity monitoring is difficult because suspicious price/volume behavior can arise from many benign mechanisms, while modern detection systems often rely on opaque models that are hard to audit and communicate. We present AIMM-X, an explainable monitoring pipeline that combines market microstructure-style signals derived from OHLCV time series with multi-source public attention signals (e.g., news and online discussion proxies) to surface time windows that merit analyst review. The system detects candidate anomalous windows using transparent thresholding and aggregation, then assigns an interpretable integrity score decomposed into a small set of additive components, allowing practitioners to trace why a window was flagged and which factors drove the score. We provide an end-to-end, reproducible implementation that downloads data, constructs attention features, builds unified panels, detects windows, computes component signals, and generates summary figures/tables. Our goal is not to label manipulation, but to provide a practical, auditable screening tool that supports downstream investigation by compliance teams, exchanges, or researchers.

An Explainable Market Integrity Monitoring System with Multi-Source Attention Signals and Transparent Scoring

TL;DR

AIMM-X tackles market integrity surveillance by enabling auditable, explainable detection using publicly available OHLCV data and multi-source attention proxies. It constructs an interpretable Integrity Score decomposed into six phi-signals, enabling transparent attribution of why a window is flagged. The study demonstrates feasibility across 24 tickers with 233 detected windows in 2024, highlighting the framework’s potential as a triage tool for compliance, research, and regulatory engagement. While not a definitive detector of manipulation, AIMM-X provides a reproducible, open framework that surface anomalies for expert review, outlines practical deployment considerations, and maps a clear path for data enhancements and validation with ground-truth cases.

Abstract

Market integrity monitoring is difficult because suspicious price/volume behavior can arise from many benign mechanisms, while modern detection systems often rely on opaque models that are hard to audit and communicate. We present AIMM-X, an explainable monitoring pipeline that combines market microstructure-style signals derived from OHLCV time series with multi-source public attention signals (e.g., news and online discussion proxies) to surface time windows that merit analyst review. The system detects candidate anomalous windows using transparent thresholding and aggregation, then assigns an interpretable integrity score decomposed into a small set of additive components, allowing practitioners to trace why a window was flagged and which factors drove the score. We provide an end-to-end, reproducible implementation that downloads data, constructs attention features, builds unified panels, detects windows, computes component signals, and generates summary figures/tables. Our goal is not to label manipulation, but to provide a practical, auditable screening tool that supports downstream investigation by compliance teams, exchanges, or researchers.
Paper Structure (73 sections, 15 equations, 2 figures, 7 tables, 1 algorithm)

This paper contains 73 sections, 15 equations, 2 figures, 7 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. The Gap in Current Surveillance Capabilities
  3. The AIMM-X Approach: Transparency Meets Practicality
  4. Motivation: Learning from Historical Detection Failures
  5. Objective.
  6. Limitation addressed.
  7. Contributions of This Work
  8. Scope and Limitations
  9. Paper Organization
  10. Related Work
  11. Market Manipulation: Theory and Detection
  12. Market Microstructure and Information Content
  13. Attention, Sentiment, and Social Media Effects
  14. Anomaly Detection and Explainable AI
  15. Reproducibility and Open Science
  16. ...and 58 more sections

Figures (2)

  • Figure 1: Distribution of composite window integrity scores $M$. The heavy-tailed distribution reflects extreme values in $\phi_1$ for early-window baseline estimation artifacts. Rank-based analysis (rank percentiles) provides more robust comparison.
  • Figure 2: Detected window counts per ticker. TLT (bond ETF) shows highest count, potentially reflecting interest rate sensitivity during 2024.