Movement Prediction Research Articles

Numerical Modelling and Prediction of Oil Slick Dispersion and Horizontal Movement at Bornholm Basin in Baltic Sea

This paper presents an original approach to predicting oil slick movement and dispersion at the water surface. Special emphasis is placed on the impact of evolving hydro-meteorological conditions and the thickness of the oil spill layer. The main gap addressed by this study lies in the need for a comprehensive understanding of how changing environmental conditions and oil thickness interact to influence the movement and dispersion of oil slicks. By focusing on this aspect, this study aims to provide valuable insights into the complex dynamics of oil spill behaviour, enhancing the ability to predict and mitigate the environmental impacts of such incidents. Self-designed software was applied to develop and modify previously established mathematical probabilistic models for predicting changes in the shape of the oil trajectory. First, a semi-Markov model of the process is constructed, and the oil thickness is analysed at the sea surface over time. Next, a stochastic-based procedure to forecast the horizontal movement and dispersion of an oil slick in diverse hydro-meteorological conditions considering a varying oil layer thickness is presented. This involves determining the trajectory and movement of a slick domain, which consists of an elliptical combination of domains undergoing temporal changes. By applying the procedure and program, a short-term forecast of the horizontal movement and dispersion of an oil slick provided its trajectory at the Bornholm Basin of the Baltic Sea within two days. The research results obtained are preliminary prediction results, although the approach considered in this paper can help responders understand the scope of the problem and mitigate the effects of environmental damage if the oil discharge reaches sensitive ecosystems. Finally, further perspectives of this research are given.

Smart Agriculture: Enhancing Security Through Animal Detection Via Deep Learning and Computer Vision

Agriculture stands as a crucial sector, making significant contributions to the economies of many countries. Nevertheless, it encounters various challenges, one of which is animal disruption. This poses a considerable threat to crops, leading to financial losses for farmers. In response to this concern, we have engineered an animal disruption warning system for agricultural settings based on YOLOv6 technology.The system operates by analyzing live video feeds from strategically placed cameras. Utilizing deep learning algorithms, it can detect and classify animals in real-time. The computer vision algorithms enable tracking and prediction of animal movements. Upon detection, the system promptly sends alerts, enabling timely and appropriate actions.In this paper, we periodically monitor the entire farm through a camera that continuously records its surroundings. The identification of animal entry is achieved using a deep learning model, and alarm systems serve as a deterrent, notifying forest officials. This report provides details on the libraries and convolutional neural networks employed in constructing the model.This research focuses on the implementation of a robust animal detection system in agricultural environments, leveraging the capabilities of deep learning. The project utilizes state-of-the-art deep neural networks and computer vision algorithms to analyze live video feeds from strategically positioned cameras across the farm. The deep learning model is trained to detect and classify various animals in real-time, contributing to the early identification of potential threats to crops.The system employs sophisticated computer vision techniques, enabling accurate tracking and prediction of animal movements within the monitored areas. Upon detection, the system triggers timely alerts, providing farmers with the necessary information to take swift and appropriate actions, thereby mitigating potential damage to crops.To achieve these objectives, the project involves periodic monitoring of the entire farm through a camera that continuously records its surroundings. The deep learning model, supported by alarm systems, effectively identifies animal entries, serving as a proactive deterrent. This research report outlines the libraries, frameworks, and convolutional neural networks employed in the development of the animal detection model, shedding light on the technical aspects of its implementation.The integration of deep learning and computer vision in agriculture not only enhances crop protection but also contributes to the sustainable and efficient management of farming practices. This research offers insights into the potential of advanced technologies to address challenges in agriculture and opens avenues for further exploration in the intersection of technology and agriculture.              

Hybrid Orthodontics for Aesthetic Deep Bite Correction—Case Series and General Clinical Considerations

Background: A range of psychosocial and aesthetic factors motivate patients to undergo orthodontic treatment. The appliance choice depends not only on the type of malocclusion, but also on the aesthetic and functional demands of the patients themselves. Nowadays, digital planning enables the manufacture of individualised and customised orthodontic appliances. However, the predictability of movements with aligner treatment has long been under discussion. This article illustrates, through a series of case reports, how a hybrid approach combining individualised aesthetic orthodontic appliances can improve the predictability of tooth movements, thereby providing additional tools for clinicians charged with choosing the best indicated and biomechanically advantageous technique. To this end, three patients with different malocclusions were treated via a hybrid approach involving clear aligners in the upper arch followed by lingual fixed appliances in the upper and lower arches. All patients were treated using ALIAS lingual PSL 0.018 × 0.018-inch slot brackets and in-house 3D-printed aligners. Findings: The hybrid approach combining aligners and fixed lingual appliances led to the successful resolution of all three complex cases in the series without prolonging treatment time. The superimpositions demonstrate the predictability of even traditionally difficult movements. In particular, the Alias PSL lingual system, used from the beginning, enabled good control of both the torque and inclination of the lower incisors. Conclusions: Combining clear aligners and fixed lingual appliances provides a highly efficient means of treating malocclusions aesthetically. In our cases, the aligners offset the lack of bite-plate effect from the lingual brackets and appliances, providing advantageous biomechanics for rotation correction and control of tip, torque and root movements. Understanding how to exploit the strengths of each appliance enables the clinician to treat adult patients efficaciously, efficiently and aesthetically.

Ground movement and settlement prediction induced by double-track curvature shield tunneling

Ground movement and settlement prediction induced by double-track curvature shield tunneling

A hierarchical deep model integrating economic facts for stock movement prediction

Accurate stock movement prediction is essential to profit from the stock market. However, this task is challenging due to the complexity and non-stationary nature of the market. Deep learning methods have obtained more attention and success in mining price movement patterns.However, some limitations affect their performances. In general, the stock market is ever-changing, and many factors affect stock movement, so capturing the stock movement patterns is hard without enough prior information. To tackle it, we consider employing economic facts to help improve the deep learning method. In this paper, we propose a novel Hierarchical Deep learning Model that fuses Economic Facts (HDMEF) to predict stock movement from the micro to the macro tiers: the individual, industry, and whole market tiers. Specifically, we present three well-designed modules to separately model them based on the Capital Asset Pricing Model (CAPM), the herding effects, and the holiday effects in the stock market. Experiments on the A-share CSI300 and CSI500 indexes demonstrate that our proposed method performs best on all test phases compared with previous competitive baselines, even an absolute improvement of 2%–3% on some test phases where all the baselines act poor, proving our method is more efficient and robust in different market conditions. In addition, we do an ablation study to analyze the role of various economic effects used in our model, and the results prove that each module is helpful for prediction.

ECHO-GL: Earnings Calls-Driven Heterogeneous Graph Learning for Stock Movement Prediction

Stock movement prediction serves an important role in quantitative trading. Despite advances in existing models that enhance stock movement prediction by incorporating stock relations, these prediction models face two limitations, i.e., constructing either insufficient or static stock relations, which fail to effectively capture the complex dynamic stock relations because such complex dynamic stock relations are influenced by various factors in the ever-changing financial market. To tackle the above limitations, we propose a novel stock movement prediction model ECHO-GL based on stock relations derived from earnings calls. ECHO-GL not only constructs comprehensive stock relations by exploiting the rich semantic information in the earnings calls but also captures the movement signals between related stocks based on multimodal and heterogeneous graph learning. Moreover, ECHO-GL customizes learnable stock stochastic processes based on the post earnings announcement drift (PEAD) phenomenon to generate the temporal stock price trajectory, which can be easily plugged into any investment strategy with different time horizons to meet investment demands. Extensive experiments on two financial datasets demonstrate the effectiveness of ECHO-GL on stock price movement prediction tasks together with high prediction accuracy and trading profitability.

Gaze-Based Interaction Adaptation for People with Involuntary Head Movements (Student Abstract)

Gaze estimation is an important research area in computer vision and machine learning. Eye-tracking and gaze-based interactions have made assistive technology (AT) more accessible to people with physical limitations. However, a non-negligible proportion of existing AT users, including those having dyskinetic cerebral palsy (CP) or severe intellectual disabilities (ID), have difficulties in using eye trackers due to their involuntary body movements. In this paper, we propose an adaptation method pertaining to head movement prediction and fixation smoothing to stabilize our target users' gaze points on the screen and improve their user experience (UX) in gaze-based interaction. Our empirical experimentation shows that our method significantly shortens the users' selection time and increases their selection accuracy.

MDGNN: Multi-Relational Dynamic Graph Neural Network for Comprehensive and Dynamic Stock Investment Prediction

The stock market is a crucial component of the financial system, but predicting the movement of stock prices is challenging due to the dynamic and intricate relations arising from various aspects such as economic indicators, financial reports, global news, and investor sentiment. Traditional sequential methods and graph-based models have been applied in stock movement prediction, but they have limitations in capturing the multifaceted and temporal influences in stock price movements. To address these challenges, the Multi-relational Dynamic Graph Neural Network (MDGNN) framework is proposed, which utilizes a discrete dynamic graph to comprehensively capture multifaceted relations among stocks and their evolution over time. The representation generated from the graph offers a complete perspective on the interrelationships among stocks and associated entities. Additionally, the power of the Transformer structure is leveraged to encode the temporal evolution of multiplex relations, providing a dynamic and effective approach to predicting stock investment. Further, our proposed MDGNN framework achieves the best performance in public datasets compared with the state-of-the-art stock investment methods.

A large-scale microblog dataset and stock movement prediction based on Supervised Contrastive Learning model

The integration of Deep Neural Networks (DNN) with Natural Language Processing (NLP) technologies has opened new avenues in financial market prediction, particularly through the utilization of textual information. This study represents a significant advancement, which offers two primary contributions to stock trend prediction: (i) the exploitation of textual data (news, comments, microblogs) using advanced DNN architectures, enhancing market information utilization; (ii) significant improvement of the accuracy of predicting the direction of stock volatility by integrating textual and neural network technologies. Meanwhile, we have crawled, filtered, and constructed a large-scale microblog dataset. This dataset includes approximately 114,992 microblog textual data from 40 Science and Technology Innovation Board (STIB) companies in China during 2021. We conducted a comprehensive analysis using various DNN techniques, including Feedback Neural Networks (FNN), Supervised Contrastive Learning (SCL), Cross Entropy (CE), and Dual Contrastive Learning (DualCL), in conjunction with bag of words models, BERT, and Roberta compilers. Our findings reveal that the SCL method, when combined with microblog data, significantly increases prediction accuracy, particularly during the COVID-19 period. Furthermore, we discovered that using a cross-stock dataset enhances the accuracy of all prediction methods, and random allocation of microblog data leads to better results than sequential allocation. Additionally, we compared the efficacy of traditional models like the CAPM, three-factor, and five-factor models against neural network-based methods. Our results suggest a notable superiority of the SCL method in increasing prediction accuracy. Finally, applying our findings to real-world trading strategies, we demonstrated the practical advantages of using the SCL method in trading, evidenced by significant improvements across all performance indicators.

Segmenting Bitcoin Transactions for Price Movement Prediction

Cryptocurrencies like Bitcoin have received substantial attention from financial exchanges. Unfortunately, arbitrage-based financial market price prediction models are ineffective for cryptocurrencies. In this paper, we utilize standard machine learning models and publicly available transaction data in blocks to predict the direction of Bitcoin price movement. We illustrate our methodology using data we merged from the Bitcoin blockchain and various online sources. This gave us the Bitcoin transaction history (block IDs, block timestamps, transaction IDs, senders’ addresses, receivers’ addresses, transaction amounts), as well as the market exchange price, for the period from 13 September 2011 to 5 May 2017. We show that segmenting publicly available transactions based on investor typology helps achieve higher prediction accuracy compared to the existing Bitcoin price movement prediction models in the literature. This transaction segmentation highlights the role of investor types in impacting financial markets. Managerially, the segmentation of financial transactions helps us understand the role of financial and cryptocurrency market participants in asset price movements. These findings provide further implications for risk management, financial regulation, and investment strategies in this new era of digital currencies.

Prediction of thermal front movement in geothermal reservoirs containing distributed natural fractures

When designing a reinjection scheme for geothermal development, one needs to predict the flow and thermal behavior between the recharge and discharge wells. Premature thermal front breakthrough is of great concern and is often observed in real fields. However, prediction methods based on homogeneous media are prone to overly optimistic estimates. In this study, geothermal reservoirs with vertical fractures which are discretely distributed in the permeable rock are considered, and the streamline simulation is developed and applied to evaluate the thermal front movement, represented by the characteristic times tD*. With 1000 simulation results for stochastically generated natural fracture systems, regression models are developed for the expected values of tD* with minimal information requirements on fracture parameters. In general, the fracture orientation θf has a stronger negative correlation with tD* than the total fracture length Σf. The intensity of natural fractures has opposing effects (delay or advance) on tD* depending on θf, and taking this effect into account improves the quality of the regression model. In addition, when the pressure difference between the doublet wells is available, the model accuracy is further improved by incorporating pressure data into the regression. The prediction models are of practical use for understanding the thermal behavior and designing a reinjection scheme in fractured geothermal reservoirs.

Decoding of unimanual and bimanual reach-and-grasp actions from EMG and IMU signals in persons with cervical spinal cord injury.

Chronic motor impairments of arms and hands as the consequence of a&#xD;cervical spinal cord injury (SCI) have a tremendous impact on activities of daily life.&#xD;A considerable number of people however retain minimal voluntary motor control in the&#xD;paralyzed parts of the upper limbs that are measurable by electromyography (EMG)&#xD;and inertial measurement units (IMUs). An integration into human-machine interfaces&#xD;(HMIs) holds promise for reliable grasp intent detection and intuitive assistive device&#xD;control.&#xD;Approach: We used a multimodal HMI incorporating EMG and IMU data to decode&#xD;reach-and-grasp movements of groups of persons with cervical SCI (n=4) and without&#xD;(control, n=13). A post-hoc evaluation of control group data aimed to identify optimal&#xD;parameters for online, co-adaptive closed-loop HMI sessions with persons with cervical&#xD;SCI. We compared the performance of real-time, Random Forest-based movement&#xD;versus rest (2 classes) and grasp type predictors (3 classes) with respect to their coadaptation&#xD;and evaluated the underlying feature importance maps.&#xD;Main results: Our multimodal approach enabled grasp decoding significantly better&#xD;than EMG or IMU data alone (p<0.05). We found the 0.25 s directly prior to the first&#xD;touch of an object to hold the most discriminative information. Our HMIs correctly&#xD;predicted 79.3 ± STD 7.4 (102.7 ± STD 2.3 control group) out of 105 trials with&#xD;grand average movement vs. rest prediction accuracies above 99.64% (100% sensitivity)&#xD;and grasp prediction accuracies of 75.39 ± STD 13.77% (97.66 ± STD 5.48% control&#xD;group). Co-adaption led to higher prediction accuracies with time, and we could&#xD;identify adaptions in feature importances unique to each participant with cervical&#xD;SCI.&#xD;Significance: Our findings foster the development of multimodal and adaptive HMIs&#xD;to allow persons with cervical SCI the intuitive control of assistive devices to improve&#xD;personal independence.

Current Research Status of Respiratory Motion for Thorax and Abdominal Treatment: A Systematic Review.

Malignant tumors have become one of the serious public health problems in human safety and health, among which the chest and abdomen diseases account for the largest proportion. Early diagnosis and treatment can effectively improve the survival rate of patients. However, respiratory motion in the chest and abdomen can lead to uncertainty in the shape, volume, and location of the tumor, making treatment of the chest and abdomen difficult. Therefore, compensation for respiratory motion is very important in clinical treatment. The purpose of this review was to discuss the research and development of respiratory movement monitoring and prediction in thoracic and abdominal surgery, as well as introduce the current research status. The integration of modern respiratory motion compensation technology with advanced sensor detection technology, medical-image-guided therapy, and artificial intelligence technology is discussed and analyzed. The future research direction of intraoperative thoracic and abdominal respiratory motion compensation should be non-invasive, non-contact, use a low dose, and involve intelligent development. The complexity of the surgical environment, the constraints on the accuracy of existing image guidance devices, and the latency of data transmission are all present technical challenges.

An Improved One-Line Evolution Formulation for the Dynamic Shoreline Planforms of Embayed Beaches

In this paper, an improved one-line evolution formulation is proposed and derived for the dynamic shoreline planforms of embayed beaches. Although embayed sandy beaches can perform several functions, serving as leisure spots and areas of coastal protection, shoreline advances and retreats occur continuously as a result of many natural forces, such as winds, waves, currents, tides, etc. The one-line evolution formulation for dynamic shoreline planforms based on the polar coordinate can be adopted to simulate high-planform-curvature shorelines and achieve better stability and simplicity in comparison with other description coordinates. While the polar coordinate and rectangular control volume are adopted to derive the one-line evolution formulation for dynamic shoreline planforms, the difference between the radial direction of the polar coordinate and the normal direction of the shoreline segment may result in inaccurate predictions of shoreline movements. In this study, a correction coefficient, which can adjust the influence of these two misaligned directions, is derived and included in the one-line evolution formulation, which is based on the polar coordinate. Thus, by considering the correction coefficient, an improved one-line evolution formulation for dynamic shoreline planforms of crenulate-shaped bays is proposed in this paper. Some numerical examples are provided to verify the merits of the proposed improved one-line evolution formulation. Moreover, the proposed numerical approach is applied to simulate the dynamic movements of the shoreline in Taitung—the southeastern part of Taiwan—and the effectiveness of the proposed formulation in solving realistic engineering applications is evidently verified.

Different ecological demands shape differences in population structure and behaviour among the two generations of the small pearl-bordered fritillary.

The population structure and behaviour of univoltine butterfly species have been studied intensively. However, much less is known about bivoltine species. In particular, in-depth studies of the differences in population structure, behaviour, and ecology between these two generations are largely lacking. Therefore, we here present a mark-release-recapture study of two successive generations of the fritillary butterfly Boloria selene performed in eastern Brandenburg (Germany). We revealed intersexual and intergenerational differences regarding behaviour, dispersal, population characteristics, and protandry. The observed population densities were higher in the second generation. The flight activity of females decreased in the second generation, but remained unchanged in males. This was further supported by the rate of wing decay. The first generation displayed a linear correlation between wing decay and passed time in both sexes, whereas the linear correlation was lost in second-generation females. The proportion of resting individuals in both sexes increased in the second generation, as well as the number of nectaring females. The choice of plant genera used for nectaring seems to be more specialised in the first and more opportunistic in the second generation. The average flight distances were generally higher for females than for males and overall higher in the first generation. Predictions of long-distance movements based on the inverse power function were also generally higher in females than in males but lower in the first generation. Additionally, we found protandry only in the first but not in the second generation, which might correlate with the different developmental pathways of the two generations. These remarkable differences between both generations might reflect an adaptation to the different ecological demands during the flight season and the different tasks they have, i.e., growth in the spring season; dispersal and colonisation of new habitats during the summer season.

Application of a Variable Weight Time Function Combined Model in Surface Subsidence Prediction in Goaf Area: A Case Study in China

To attain precise forecasts of surface displacements and deformations in goaf areas (a void or cavity that remains underground after the extraction of mineral resources) following coal extraction, this study based on the limitations of individual time function models, conducted a thorough analysis of how the parameters of the model impact subsidence curves. Parameter estimation was conducted using the trust-region reflective algorithm (TRF), and the time function models were identified. Then we utilized a combined model approach and introduced the sliding window mechanism to assign variable weights to the model. Based on this, the combined model was used for prediction, followed by the application of this composite prediction to engineering scenarios for the dynamic forecasting of surface movements and deformations. The results indicated that, in comparison with DE, GA, PSO algorithms, the TRF exhibited superior stability and convergence. The parameter models obtained using this method demonstrated a higher level of predictive accuracy. Moreover, the predictive precision of the variable-weight time function combined model surpassed that of corresponding individual time function models. When employing six different variable-weight combination prediction models for point C22, the Weibull-MMF model demonstrated the most favorable fitting performance, featuring a root mean square error (RMSE) of 32.98 mm, a mean absolute error (MAE) of 25.66 mm, a mean absolute percentage error (MAPE) of 7.67%; the correlation coefficient R2 reached 0.99937. These metrics consistently outperformed their respective individual time function models. Additionally, in the validation process of the combined model at point C16, the residuals were notably smaller than those of individual models. This reaffirmed the accuracy and reliability of the proposed variable-weight combined model. Given that the variable-weight combination model was an evolution from individual time function models, its applicability extends to a broader range, offering valuable guidance for the dynamic prediction of surface movement and deformation in mining areas.

Strategic Content Generation and Monetization in Financial Social Media

Strategic Content Generation and Monetization in Financial Social Media Abstract Financial social media, which relies on social media analysts (SMAs) to contribute content to investors, is a crucial channel for investors to gain financial information and for SMAs to monetize their content. The interactive nature of financial social media has given SMAs the opportunity to gain access to the investor preferences of their own audience base for financial content. Our study documents that SMAs would exploit this opportunity to strategically generate and monetize content by catering to investor preferences. Specifically, SMAs would increase the (negative) sentiment of the content if paid subscribers’ preferences for (negative) sentiment grow. Additionally, an SMA is more likely to produce paid content when the expected free readership increases and is less likely to do so when the expected paid subscriptions increase. Our findings suggest that the sentiment of financial social media content is not a mere reflection or prediction of stock market movements but also a result of SMAs’ reaction to investor preferences. We thus illustrate an approach to identify the SMAs who may amplify the investors’ confirmation biases because of such catering behaviors so that platform managers and regulators alike can utilize this method to improve the content quality of financial social media.

Southbank Place development: tunnel movement predictions and monitoring performance

The Southbank Place project comprised the redevelopment of the UK headquarters of the Shell Company (formerly known as the Shell Centre) in the Waterloo area of London. The project included the demolition of all structures within the site except for the iconic tower building, local extension of the existing basement and the construction of eight new mixed-use buildings. London Underground's (LU) Bakerloo Line and Northern Line running tunnels pass beneath the site. The excavation of the basement and construction of the Shell Centre in the late 1950s caused movements of the underlying Bakerloo Line tunnels that were much larger than expected. The monitoring of these tunnels between 1958 and 1995 has provided a classic case history for the behaviour of London Clay. The proposed redevelopment offered an opportunity to analyse and monitor the behaviour of LU tunnels over a period of 60 years and two phases of significant construction. This paper presents a back-analysis of the historical monitoring of the Bakerloo Line tunnels and the calibration of a sophisticated finite-element model. The model was subsequently used to predict the impact of the redevelopment on the tunnels as part of the assurance procedure with LU.

Nonlinear multi independent variables in quantifying river bank erosion using Neural Network AutoRegressive eXogenous (NNARX) model

This study proposed a novel application of Neural Network AutoRegressive eXogenous (NNARX) model in predicting nonlinear behaviour of riverbank erosion rates which is difficult to be achieved with good accuracy using conventional approaches. This model can estimate complex river bank erosion rates with flow variations. The NNARX model analysed to a set of primary data, 60% (203 data for training) and 40% (135 data for testing), which were collected from Sg. Bernam, Malaysia. A set of nondimensional parameters, known as functional relationship, used as an input to the NNARX model has been established using the method of repeating variables. The One-Step-Ahead time series prediction plots are used to assess the accuracy of all developed models. Model no. 6 (5 independent variables with 10 hidden layers) gives good predictive performance, supported by the graphical analysis with discrepancy ratio of 94% and 90% for training and testing datasets. This finding is consistent with model accuracy result, where Model no. 6 achieved R2 of 0.932 and 0.788 for training and testing datasets, respectively. Result shows that bank erosion is maximized when the near-bank velocity between 0.2 and 0.5 m/s, and the riverbank erosion is between 1.5 and 1.8 m/year. On the other hand, higher velocities ranging from 0.8 to 1.3 m/s induces erosion at a rate between 0.1 and 0.4 m/year. Sensitivity analysis shows that the highest accuracy of 91% is given by the ratio of shear velocity to near-bank velocity followed by boundary shear stress to near-bank velocity ratio (88.5%) and critical shear stress to near-bank velocity ratio (88.2%). It is concluded that the developed model has accurately predicted non-linear behaviour of riverbank erosion rates with flow variations. The study's findings provide valuable insights in advanced simulations and predictions of channel migration, encompassing both lateral and vertical movements, the repercussions on the adjacent river corridor, assessing the extent of land degradation and in formulating plans for effective riverbank protection and management measures.

Stock price forecasting in Indonesia stock exchange using deep learning: a comparative study

In 2022, the Indonesia stock exchange (IDX) listed 825 companies, making it challenging to identify low-risk companies. Stock price forecasting and price movement prediction are vital issues in financial works. Deep learning has previously been implemented for stock market analysis, with promising results. Because of the differences in architecture and stock issuers in each study report, a consensus on the best stock price forecasting model has yet to be reached. We present a methodology for comparing the performance of convolutional neural networks (CNN), gated recurrent units (GRU), long short-term memory (LSTM), and graph convolutional networks (GCN) layers. The four layers types combination yields 11 architectures with two layers stacked maximum, and the architectures are performance compared in stock price predicting. The dataset consists of open, highest, lowest, closed price, and volume transactions and has 2,588,451 rows from 727 companies in IDX. The best performance architecture was chosen by a vote based on the coefficient of determination (R2), mean squared error (MSE), root mean square error (RMSE), mean absolute percent error (MAPE), and f1-score. TFGRU is the best architecture, producing the finest results on 315 companies with an average score of RMSE is 553.327, MAPE is 0.858, and f1-score is 0.456.