Stock Market Prediction Based on Generative Adversarial Network

Zhang, Kang; Zhong, Guoqiang; Dong, Junyu; Wang, Shengke; Wang, Yong

doi:10.1016/j.procs.2019.01.256

Cited by 235 publications

(131 citation statements)

References 8 publications

Supporting

Mentioning

107

Contrasting

Order By: Relevance

“…Most importantly, the question regarding a reasonable network architecture has to be answered. There exist promising approaches in the literature aiming at applying the GAN idea to time series [12], recently also with a focus on financial time series, see [26,37,42,43]. However, our numerical analysis suggests that the architectures do not work as well as the proposed Quant GAN model constructed in the present paper.…”

Section: Introductionmentioning

confidence: 71%

Quant GANs: deep generation of financial time series

Wiese

Knobloch

Korn

et al. 2020

Quantitative Finance

182

113

View full text Add to dashboard Cite

Modeling financial time series by stochastic processes is a challenging task and a central area of research in financial mathematics. In this paper, we break through this barrier and present Quant GANs, a data-driven model which is inspired by the recent success of generative adversarial networks (GANs). Quant GANs consist of a generator and discriminator function which utilize temporal convolutional networks (TCNs) and thereby achieve to capture longer-ranging dependencies such as the presence of volatility clusters. Furthermore, the generator function is explicitly constructed such that the induced stochastic process allows a transition to its risk-neutral distribution. Our numerical results highlight that distributional properties for small and large lags are in an excellent agreement and dependence properties such as volatility clusters, leverage effects, and serial autocorrelations can be generated by the generator function of Quant GANs, demonstrably in high fidelity.

show abstract

Section: Introductionmentioning

confidence: 71%

Quant GANs: deep generation of financial time series

Wiese

Knobloch

Korn

et al. 2020

Quantitative Finance

182

113

View full text Add to dashboard Cite

show abstract

“…On the other side, Autoregressive Integrated Moving Average (ARIMA) is one of the popular models when comes to the time series prediction. In Zhang"s research, he and his team have selected the GAN model with Long-Short Term Memory (LSTM) network as the generator and Multi-Layer Perceptron (MLP) as the discriminator for predicting the stock price [11].…”

Section: Neural Graph Collaborative Filtering Frameworkmentioning

confidence: 99%

“…On the other hand, the Discriminator"s role is to distinguish real and fake data, input from both real datasets and generator sets. Both algorithms are trained together until it reaches a stage that the generator is capable of generating the fake data that the discriminator unable to classify as fake input [11]. In summary, the Discriminator function is to maximize the probability of identifying correct input while the Generator is continuously trained to minimize the probability of letting the Discriminator identified as the generated output as fake input [20].…”

Section: Generative Adversarial Network (Gan)mentioning

confidence: 99%

Predicting Book Sales Trend using Deep Learning Framework

Feng¹,

Choy²,

Laik³

2020

IJACSA

View full text Add to dashboard Cite

A deep learning framework like Generative Adversarial Network (GAN) has gained popularity in recent years for handling many different computer visions related problems. In this research, instead of focusing on generating the near-real images using GAN, the aim is to develop a comprehensive GAN framework for book sales ranks prediction, based on the historical sales rankings and different attributes collected from the Amazon site. Different analysis stages have been conducted in the research. In this research, a comprehensive data preprocessing is required before the modeling and evaluation. Extensive predevelopment on the data, related features selections for predicting the sales rankings, and several data transformation techniques are being applied before generating the models. Later then various models are being trained and evaluated on prediction results. In the GAN architecture, the generator network that used to generate the features is being built, and the discriminator network that used to differentiate between real and fake features is being trained before the predictions. Lastly, the regression GAN model prediction results are compared against the different neural network models like multilayer perceptron, deep belief network, convolution neural network.

show abstract

“…However plenty of different other methods based on computational intelligence accomplish accurate predictions on stock market (e.g. [1][2][3][4]). These algorithms include evolutionary computation through the genetic algorithm to support vector machine or various neural network-based approach [5,6], deep learning methods such as deep belief net coupled with multi-linear perceptron [7][8][9][10].…”

Section: Introductionmentioning

confidence: 99%

Construction of Confidence Interval for a Univariate Stock Price Signal Predicted Through Long Short Term Memory Network

Dey

Gouda

2020

Ann. Data. Sci.

View full text Add to dashboard Cite

In this paper, we show an innovative way to construct bootstrap confidence interval of a signal estimated based on a univariate LSTM model. We take three different types of bootstrap methods for dependent set up. We prescribe some useful suggestions to select the optimal block length while performing the bootstrapping of the sample. We also propose a benchmark to compare the confidence interval measured through different bootstrap strategies. We illustrate the experimental results through some stock price data set.

show abstract

Stock Market Prediction Based on Generative Adversarial Network

Cited by 235 publications

References 8 publications

Quant GANs: deep generation of financial time series

Quant GANs: deep generation of financial time series

Predicting Book Sales Trend using Deep Learning Framework

Construction of Confidence Interval for a Univariate Stock Price Signal Predicted Through Long Short Term Memory Network

Contact Info

Product

Resources

About