Restricted Boltzmann Machines for the Prediction of Trends in Financial Time Series

Assis, Carlos A. S.; Pereira, Adriano C. M.; Carrano, Eduardo G.; Ramos, Rafael; Dias, Wanderson

doi:10.1109/ijcnn.2018.8489163

Cited by 16 publications

(9 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As for predictors, some trading indicators such as opening price and turnover rate are widely used to predict the financial indices. Many studies use the historical price to predict the trend of the stock market (Assis et al, 2018;Chen et al, 2019). Al-Thelaya et al (2019) augment the predictors into some technical indicators.…”

Section: Data and Methodology Datamentioning

confidence: 99%

Green Bond Index Prediction Based on CEEMDAN-LSTM

2022

View full text Add to dashboard Cite

Green bonds, which are designed to finance for environment-friendly or sustainable projects, have attracted more and more investors’ attention. However, the study in this field is still relatively limited, especially in forecasting the market’s future trends. In this paper, a hybrid model combining CEEMDAN and LSTM is introduced to predict green bond market in China (represented by CUFE-CNI High Grade Green Bond Index). In order to evaluate the performance of our model, we also use EMD to decompose the green bond index. Our empirical result suggests that, compared with EMD-LSTM and LSTM models, CEEMDAN-LSTM is the most accurate model in green bond index forecasting. Meanwhile, we find that indices from the crude oil market and green stock market are both effective predictors, which also provides ground on the correlations between the green bond market and other financial markets.

show abstract

Section: Data and Methodology Datamentioning

confidence: 99%

Green Bond Index Prediction Based on CEEMDAN-LSTM

2022

View full text Add to dashboard Cite

show abstract

“…While some methods depend on edge features, texture descriptors, and shape contexts (de Campos et al, ), others use probabilistic models (Weinman, Learned‐Miller, & Hanson, ) (Weinman, Learned‐Miller, & Hanson, ) (Fan & Fan, ). Many algorithms learn features from unlabelled data (Hinton, Osindero, & Teh, ) (Assis, Pereira, Carrano, Ramos, & Dias, ) (Ranzato, Krizhevsky, & Hinton, ) (Raina, Battle, Lee, Packer, & Ng, ) that specifically lead them to recognize characters in the images studied in the context of scanned documents and books (Nagy, ) (Mori, Suen, & Yamamoto, ). Others recognize currency serial num‐ bers (Qian, Qian, & Zhang, ).…”

Section: Related Workmentioning

confidence: 99%

Blockchain for tracking serial numbers in money exchanges

Mohamed

Aziz

Mohamed

et al. 2019

Intell Sys Acc Fin Mgmt

View full text Add to dashboard Cite

Money exchange is one of the most common day-to-day activities performed by humans in the daily market. This paper presents an approach to money tracking through a blockchain. The proposed approach consists of three main components: serial number localization, serial number recognition, and a blockchain to store all transactions and ownership transfers. The approach was tested with a total of 110 banknotes of different currency types and achieved an average accuracy of 91.17%. We conducted a user study in real-time with 21 users, and the mean accuracy across all users was 86.42%. Each user gave us feedback on the proposed approach, and most of them welcomed the idea. K E Y W O R D Simage processing, banknote serial number recognition (SNR), blockchain

show abstract

“…This setting can be formulated as discriminative as well as generative learning task (Liu and Webb 2010). Applications have been studied in a large variety of domains such as the analysis of quantum many-body systems, statistics, biochemistry, social networks, signal processing, and finance; see, e.g., (Torlai et al 2018;Carleo et al 2018;Carleo and Troyer 2017;Nomura et al 2017;Anshu et al 2020;Melko et al 2019;Hrasko et al 2015;Tubiana et al 2019;Liu et al 2013;Mohamed and Hinton 2010;Assis et al 2018). However, BMs are complicated to train in practice because the loss function's derivative requires the evaluation of a normalization factor, the partition function, that is generally difficult to compute.…”

Section: Introductionmentioning

confidence: 99%

Variational quantum Boltzmann machines

Zoufal

Lucchi

Woerner

2021

Quantum Mach. Intell.

View full text Add to dashboard Cite

This work presents a novel realization approach to quantum Boltzmann machines (QBMs). The preparation of the required Gibbs states, as well as the evaluation of the loss function’s analytic gradient, is based on variational quantum imaginary time evolution, a technique that is typically used for ground-state computation. In contrast to existing methods, this implementation facilitates near-term compatible QBM training with gradients of the actual loss function for arbitrary parameterized Hamiltonians which do not necessarily have to be fully visible but may also include hidden units. The variational Gibbs state approximation is demonstrated with numerical simulations and experiments run on real quantum hardware provided by IBM Quantum. Furthermore, we illustrate the application of this variational QBM approach to generative and discriminative learning tasks using numerical simulation.

show abstract

Restricted Boltzmann Machines for the Prediction of Trends in Financial Time Series

Cited by 16 publications

References 17 publications

Green Bond Index Prediction Based on CEEMDAN-LSTM

Green Bond Index Prediction Based on CEEMDAN-LSTM

Blockchain for tracking serial numbers in money exchanges

Variational quantum Boltzmann machines

Contact Info

Product

Resources

About