Blockchain, DLT and Fintech
Loading...
Paper Type
Complete
Paper Number
1143
Description
Stock forecast with candlestick patterns is heavily based on template-oriented and rule-based heuristics, which requires laborious sample labelling and profound financial expertise. These methods are retrospective in nature and fail to capture premature or partial signals in candlesticks. Such rigidity limits the application of candlesticks primarily to classification tasks. Thus, we propose a novel, end-to-end deep learning model, GANStick, to address all these issues. GANStick is a conditional DCGAN-convolutional BiLSTMbased model which generates future predictive candlesticks to augment multistep time series forecasting with regression. GANStick has been empirically shown to significantly beat multiple baseline implementations, with an average error rate of 68% lower across all five timesteps on the dataset composed of 11 large-cap US stocks. GANStick is the first work in automating the workflow from candlestick pattern recognition and generation to quantifying future price volatility, with the novel generative candlestick approach using the generative adversarial network.
Recommended Citation
Wong, Man Hing; LEE, Lik Hang; and Hui, Pan, "GANStick: US Stock Forecasting with GAN-Generated Candlesticks" (2020). ICIS 2020 Proceedings. 1.
https://aisel.aisnet.org/icis2020/blockchain_fintech/blockchain_fintech/1
GANStick: US Stock Forecasting with GAN-Generated Candlesticks
Stock forecast with candlestick patterns is heavily based on template-oriented and rule-based heuristics, which requires laborious sample labelling and profound financial expertise. These methods are retrospective in nature and fail to capture premature or partial signals in candlesticks. Such rigidity limits the application of candlesticks primarily to classification tasks. Thus, we propose a novel, end-to-end deep learning model, GANStick, to address all these issues. GANStick is a conditional DCGAN-convolutional BiLSTMbased model which generates future predictive candlesticks to augment multistep time series forecasting with regression. GANStick has been empirically shown to significantly beat multiple baseline implementations, with an average error rate of 68% lower across all five timesteps on the dataset composed of 11 large-cap US stocks. GANStick is the first work in automating the workflow from candlestick pattern recognition and generation to quantifying future price volatility, with the novel generative candlestick approach using the generative adversarial network.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.