Long-Term Modeling of Financial Machine Learning for Active Portfolio Management

Paper: arXiv 2301.12346

Authors: Kazuki Amagai, Tomoya Suzuki

Abstract

In the practical business of asset management by investment trusts and the like, the general practice is to manage over the medium to long term owing to the burden of operations and increase in transaction costs with the increase in turnover ratio. However, when machine learning is used to construct a management model, the number of learning data decreases with the increase in the long-term time scale; this causes a decline in the learning precision. Accordingly, in this study, data augmentation was applied by the combined use of not only the time scales of the target tasks but also the learning data of shorter term time scales, demonstrating that degradation of the generalization performance can be inhibited even if the target tasks of machine learning have long-term time scales. Moreover, as an illustration of how this data augmentation can be applied, we conducted portfolio management in which machine learning of a multifactor model was done by an autoencoder and mispricing was used from the estimated theoretical values. The effectiveness could be confirmed in not only the stock market but also the FX market, and a general-purpose management model could be constructed in various financial markets.

Complexity vs Empirical Score

• Math Complexity: 5.5/10
• Empirical Rigor: 6.0/10 (capped from 6.5: transaction costs not modelled; see penalty flags)
• Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper introduces a clever data augmentation strategy combining short-term and long-term time scales to overcome data scarcity in long-horizon machine learning models. It demonstrates practical application using autoencoder-based multifactor models across both equity and foreign exchange markets with solid empirical validation.

Research Flowchart

  flowchart TD
    A[Goal: Improve Machine Learning for Long-Term Portfolio Management] --> B[Data Inputs: Stock and FX Markets Data]
    B --> C[Method: Data Augmentation combining Short-Term & Long-Term Scales]
    C --> D[Model: Autoencoder-based Multifactor Model]
    D --> E[Process: Estimate Theoretical Values & Identify Mispricing]
    E --> F[Outcomes: Inhibited Generalization Degradation & Effective Multi-Market Strategy]