Analysis of Gradient Descent Optimization Techniques with Gated Recurrent Unit for Stock Price Prediction: A Case Study on Banking Sector of Nepal Stock Exchange

Saud, Arjun Singh; Shakya, Subarna

doi:10.3126/jist.v24i2.27247

Cited by 7 publications

(3 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although the GRU model achieved an encouraging prediction accuracy, it has some drawbacks. One drawback is that the GRU uses the stochastic gradient descent optimization algorithm to update weights, which risks falling into local optimization [60]. Another drawback is that the deep GRU model demands a larger dataset size than conventional ML models [10].…”

Section: Discussionmentioning

confidence: 99%

Using Complementary Ensemble Empirical Mode Decomposition and Gated Recurrent Unit to Predict Landslide Displacements in Dam Reservoir

Yang

Xiao

Huang³

2022

Sensors

View full text Add to dashboard Cite

It is crucial to predict landslide displacement accurately for establishing a reliable early warning system. Such a requirement is more urgent for landslides in the reservoir area. The main reason is that an inaccurate prediction can lead to riverine disasters and secondary surge disasters. Machine learning (ML) methods have been developed and commonly applied in landslide displacement prediction because of their powerful nonlinear processing ability. Recently, deep ML methods have become popular, as they can deal with more complicated problems than conventional ML methods. However, it is usually not easy to obtain a well-trained deep ML model, as many hyperparameters need to be trained. In this paper, a deep ML method—the gated recurrent unit (GRU)—with the advantages of a powerful prediction ability and fewer hyperparameters, was applied to forecast landslide displacement in the dam reservoir. The accumulated displacement was firstly decomposed into a trend term, a periodic term, and a stochastic term by complementary ensemble empirical mode decomposition (CEEMD). A univariate GRU model and a multivariable GRU model were employed to forecast trend and stochastic displacements, respectively. A multivariable GRU model was applied to predict periodic displacement, and another two popular ML methods—long short-term memory neural networks (LSTM) and random forest (RF)—were used for comparison. Precipitation, reservoir level, and previous displacement were considered to be candidate-triggering factors for inputs of the models. The Baijiabao landslide, located in the Three Gorges Reservoir Area (TGRA), was taken as a case study to test the prediction ability of the model. The results demonstrated that the GRU algorithm provided the most encouraging results. Such a satisfactory prediction accuracy of the GRU algorithm depends on its ability to fully use the historical information while having fewer hyperparameters to train. It is concluded that the proposed model can be a valuable tool for predicting the displacements of landslides in the TGRA and other dam reservoirs.

show abstract

Section: Discussionmentioning

confidence: 99%

Using Complementary Ensemble Empirical Mode Decomposition and Gated Recurrent Unit to Predict Landslide Displacements in Dam Reservoir

Yang

Xiao

Huang³

2022

Sensors

View full text Add to dashboard Cite

show abstract

“…These selected features are then used as input features to train our two-layered GRU model, which is then used for forecasting stock prices. The related hyper parameters, like the number of epochs, choice of the optimizer, batch size, and time step, are likewise appropriately described [34,35]. Lastly, we calculated the correctness of our model to verify the results, for Binary PSO number of iterations used is 10, while only 1 iteration was used for Binary GWO.…”

Section: A Aggregation Methodsmentioning

confidence: 99%

Aggregation and Hybrid Approaches Using Bpso and Bgwo for Feature Selection for Indian Stock Index Price Prediction Using Gru

Gupta,

Singh Jadon

2023

IJEAST

View full text Add to dashboard Cite

The stock exchange is a significant component of the economy, and forecasting its development is essential. Several deep-learning time series models based on RNN and its variants are used to forecast the stock market, but their accuracy still needs to be improved. Optimization strategies, such as GA, PSO, GWO, and others, have been used to enhance the reliability of these models. In this proposed paper, we have used the Binary GWO and Binary PSO algorithms to optimize the input characteristics of a two-layer GRU model for forecasting Indian stock index price. To increase the accuracy of these models, we have suggested two distinct approaches: Aggregation Method and the Hybrid method. In the 10- year time frame, the data indicated that the aggregation approach produced superior accuracy than standard Binary GWO and Binary PSO independently. Although the suggested hybrid method did not out perform individual Binary PSO methods in terms of accuracy, it performed better than the Binary GWO model in a 10- year time frame. Moreover, it showed promise for future improvement. The proposed approaches have the potential to make essential contributions to the index of stock price forecasting, particularly in the Indian market.

show abstract

“…Besides, Adam optimizer was used with the network so that the learning rate can adapt itself. It is the best optimization strategy for stock price prediction (Saud & Shakya, 2019). The GRU network was trained using a batch gradient descent approach with a batch size of 32.…”

Section: Configuration Of Gru Networkmentioning

confidence: 99%

Analysis of L2 Regularization Hyper Parameter for Stock Price Prediction

Saud

Shakya

2021

J. Inst. Sci. Tech.

Self Cite

View full text Add to dashboard Cite

Nowadays stock price prediction is an active area of research among machine learning researchers. One of the main problems with machine learning models is overfitting. Regularization techniques are widely used approaches to avoid over-fitted models. L2 regularization is one of the most popular and widely used regularization techniques. Regularization hyperparameter (ʎ) is one key parameter to be optimized for a well-generalized machine learning model. Hyperparameters can’t be learned by machine learning models during the learning process. We need to find their optimal value through experiments. This research work analyzed the L2 regularization hyperparameter used with a gated recurrent unit (GRU) network for stock price prediction. We experimented with five stocks from the Nepal Stock Exchange (NEPSE) and observed that stock price can be predicted with lower mean squared errors (MSEs) when the value of ʎ was around 0.0005. Therefore, this research paper recommended using ʎ=0.0005 with L2 regularization for stock price prediction.

show abstract

Analysis of Gradient Descent Optimization Techniques with Gated Recurrent Unit for Stock Price Prediction: A Case Study on Banking Sector of Nepal Stock Exchange

Cited by 7 publications

References 7 publications

Using Complementary Ensemble Empirical Mode Decomposition and Gated Recurrent Unit to Predict Landslide Displacements in Dam Reservoir

Using Complementary Ensemble Empirical Mode Decomposition and Gated Recurrent Unit to Predict Landslide Displacements in Dam Reservoir

Aggregation and Hybrid Approaches Using Bpso and Bgwo for Feature Selection for Indian Stock Index Price Prediction Using Gru

Analysis of L2 Regularization Hyper Parameter for Stock Price Prediction

Contact Info

Product

Resources

About