CIMTS: candlestick-integrated multi-modal trading system for stock indices

Juhyeon Jang; Jaeyun Kim

doi:10.7717/peerj-cs.4069

ResearchHub

CIMTS: candlestick-integrated multi-modal trading system for stock indices - ResearchHub

Verified recordOpen access

CIMTS: candlestick-integrated multi-modal trading system for stock indices

Juhyeon Jang, Jaeyun Kim

PeerJ Computer Science2026doi:10.7717/peerj-cs.4069

Citation graph

Abstract

Stock price analysis based solely on technical indicators, such as in traditional stock price prediction studies, relies heavily on observations from individual sample markets and fails to consider insights from other modalities, making it imperative to study the implicit interdependence between different modalities. A Candlestick-Integrated Multi-Modal Trading System (CIMTS) is proposed that integrates 1D numerical time-series with 2D spatial candlestick chart-image data. The system analyzes multimodal data by employing Long Short-Term Memory and Vision Transformer models in parallel, identifying graphical signals of stock prices using a late fusion approach to merge the outputs from the two models. The system was tested over an out-of-sample period of approximately 3 years, partitioned from a 12-year dataset of key US and Korean stock indices, including the NASDAQ Composite, S&P500, KOSPI, and KOSDAQ, under different market conditions. The experimental results showed that the CIMTS outperformed single-modality models in terms of aggressiveness ( i.e ., higher cumulative returns and a greater number of trades) and stability ( i.e ., lower maximum drawdown and consistent win rates). In particular, the CIMTS outperformed the labeling method that considers next-day movements in the Korean market and the 5-day consecutive trend in the US market. The proposed system achieved cumulative returns of 42.52%, 51.37%, 39.36%, and 95.47% for NASDAQ, S&P500, KOSPI, and KOSDAQ, respectively (payoff ratios: 1.023, 0.877, 1.166, and 1.195; profit factors: 1.364, 1.638, 1.306, and 1.890, respectively), demonstrating consistently improved trading performance over the Buy & Hold benchmark across all tested market regimes.

Authors

Juhyeon Jang#1 Jaeyun Kim#2

Institutions

No ROR-resolved institution is linked to this work yet.

Provenance

crossref

Confidence 100%

unpaywall

Confidence 95%

doaj

Confidence 92%

datacite

Confidence 0%

Referenced works

View all references →

No local reference links have been materialized yet.

Cited by

View all cited-by records →

No local citing links have been materialized yet.

External reference records

Stock price prediction using the ARIMA model

10.1109/uksim.2014.67 · 2014

A dynamic trading rule based on filtered flag pattern recognition for stock market price forecasting

10.1016/j.eswa.2017.03.028 · 2017

Multimodal machine learning: a survey and taxonomy

10.1109/tpami.2018.2798607 · 2018

LSTM based decision support system for swing trading in stock market

10.1016/j.knosys.2021.107994 · 2022

Wrapper ANFIS-ICA method to do stock market timing and feature selection on the basis of Japanese Candlestick

10.1016/j.eswa.2015.08.010 · 2015

Integrating StockTwits with sentiment analysis for better prediction of stock price movement

10.1109/icomet.2018.8346382 · 2018

Using neural networks and technical indicators for generating stock trading signals

2002

Meltdown modelling: could agent-based computer models prevent another financial crisis?

10.1038/460680a · 2009

Shortlisting machine learning-based stock trading recommendations using candlestick pattern recognition

10.1016/j.eswa.2022.119493 · 2023

Additional external references

PreBit—a multimodal model with Twitter FinBERT embeddings for extreme price movement prediction of Bitcoin

10.1016/j.eswa.2023.120838 · doi-reference

Prediction of stock price direction using a hybrid GA-XGBoost algorithm with a three-stage feature engineering process

10.1016/j.eswa.2021.115716 · doi-reference

FinMultiTime: a four-modal bilingual dataset for financial time-series analysis

10.48550/arxiv.2506.05019 · doi-reference

Stock movement prediction via gated recurrent unit network based on reinforcement learning with incorporated attention mechanisms

10.1016/j.neucom.2021.09.072 · doi-reference

Predicting short-term stock prices using ensemble methods and online data sources

10.1016/j.eswa.2018.06.016 · doi-reference

Stock market index prediction based on reservoir computing models

10.1016/j.eswa.2021.115022 · doi-reference

A method for automatic stock trading combining technical analysis and nearest neighbor classification

10.1016/j.eswa.2010.03.033 · doi-reference

Forecasting trends of high-frequency KOSPI200 index data using learning classifiers

10.1016/j.eswa.2012.04.015 · doi-reference

Very deep convolutional networks for large-scale image recognition

10.48550/arxiv.1409.1556 · doi-reference

Forecasting price movements using technical indicators: investigating the impact of varying input window length

10.1016/j.neucom.2016.11.095 · doi-reference

Deep convolutional neural networks for computer-aided detection: CNN architectures, dataset characteristics and transfer learning

10.1109/tmi.2016.2528162 · doi-reference

GPM: a graph convolutional network based reinforcement learning framework for portfolio management

10.1016/j.neucom.2022.04.105 · doi-reference

Stock price crash prediction based on multimodal data machine learning models

10.1016/j.frl.2024.105195 · doi-reference

ImageNet large scale visual recognition challenge

10.1007/s11263-015-0816-y · doi-reference

Stock price prediction using deep learning and frequency decomposition

10.1016/j.eswa.2020.114332 · doi-reference

Recurrent neural network and a hybrid model for prediction of stock returns

10.1016/j.eswa.2014.12.003 · doi-reference

Short-term stock trends prediction based on sentiment analysis and machine learning

10.1007/s00500-021-06602-7 · doi-reference

Feature selection and deep neural networks for stock price direction forecasting using technical analysis indicators

10.1016/j.mlwa.2021.100060 · doi-reference

Predicting stock and stock price index movement using trend deterministic data preparation and machine learning techniques

10.1016/j.eswa.2014.07.040 · doi-reference

Time series data analysis of stock price movement using machine learning techniques

10.1007/s00500-020-04957-x · doi-reference