A machine learning approach to support decision in insider trading detection

Paper: arXiv 2212.05912

Authors: Piero Mazzarisi, Adele Ravagnani, Paola Deriu, Fabrizio Lillo, Francesca Medda, Antonio Russo

Abstract

Identifying market abuse activity from data on investors’ trading activity is very challenging both for the data volume and for the low signal to noise ratio. Here we propose two complementary unsupervised machine learning methods to support market surveillance aimed at identifying potential insider trading activities. The first one uses clustering to identify, in the vicinity of a price sensitive event such as a takeover bid, discontinuities in the trading activity of an investor with respect to his/her own past trading history and on the present trading activity of his/her peers. The second unsupervised approach aims at identifying (small) groups of investors that act coherently around price sensitive events, pointing to potential insider rings, i.e. a group of synchronised traders displaying strong directional trading in rewarding position in a period before the price sensitive event. As a case study, we apply our methods to investor resolved data of Italian stocks around takeover bids.

Complexity vs Empirical Score

• Math Complexity: 3.5/10
• Empirical Rigor: 6.5/10
• Quadrant: Street Traders — practical and empirical, lighter on theory

Why this score: The paper introduces two practical unsupervised machine learning frameworks to tackle the difficult domain of insider trading and ring detection using granular Italian transaction data. While the underlying mathematical formalisms are light, relying mostly on standard clustering and network/synchrony metrics, the empirical application to investor-resolved data around takeover bids provides strong practical relevance for market surveillance.

Research Flowchart

  flowchart TD
    A[Goal: Detect Insider Trading in High-Noise Equity Data] --> B[Input Data: Italian Stock Investor-Resolved Data around Takeover Bids]
    
    B --> C[Unsupervised Machine Learning Methodology]
    
    C --> D[Method 1: Clustering Analysis]
    C --> E[Method 2: Coherent Group Detection]
    
    D --> F[Identifies Individual Discontinuities vs. Past History & Peers]
    E --> G[Identifies Potential Insider Rings & Synchronized Traders]
    
    F --> H[Outcomes: Enhanced Market Surveillance & Decision Support]
    G --> H