Data Augmentation for Deep Candlestick Learner

TL;DR

This paper introduces a novel data augmentation method for candlestick data in financial trading, enabling effective deep learning models with limited labeled data by generating high-quality, human-indistinguishable samples.

Contribution

Proposes a Modified Local Search Attack Sampling method specifically for augmenting candlestick data, addressing the scarcity of augmentation techniques in finance.

Findings

Generated data are hard to distinguish from real data by humans.

Method improves deep learning performance on small datasets.

Opens new avenues for applying machine learning in finance.

Abstract

To successfully build a deep learning model, it will need a large amount of labeled data. However, labeled data are hard to collect in many use cases. To tackle this problem, a bunch of data augmentation methods have been introduced recently and have demonstrated successful results in computer vision, natural language and so on. For financial trading data, to our best knowledge, successful data augmentation framework has rarely been studied. Here we propose a Modified Local Search Attack Sampling method to augment the candlestick data, which is a very important tool for professional trader. Our results show that the proposed method can generate high-quality data which are hard to distinguish by human and will open a new way for finance community to employ existing machine learning techniques even if the dataset is small.