Research on Stock Price Prediction Method Based on the GAN-LSTM-Attention Model

Stock price prediction is a typical complex time series prediction problem characterized by dynamics,nonlinearity,and complexity.This paper introduces a generative adversarial network model that incorporates an attention mechanism(GAN-LSTM-Attention)to improve the accuracy of stock price prediction.Firstly,the generator of this model combines the Long and Short-Term Memory Network(LSTM),the Attention Mechanism and,the Fully-Connected Layer,focusing on generating the predicted stock price.The discriminator combines the Convolutional Neural Network(CNN)and the Fully-Connected Layer to discriminate between real stock prices and generated stock prices.Secondly,to evaluate the practical application ability and generalization ability of the GAN-LSTM-Attention model,four representative stocks in the United States of America(USA)stock market,namely,Standard&Poor’s 500 Index stock,Apple Incorporatedstock,AdvancedMicroDevices Incorporatedstock,and Google Incorporated stock were selected for prediction experiments,and the prediction performance was comprehensively evaluated by using the three evaluation metrics,namely,mean absolute error(MAE),root mean square error(RMSE),and coefficient of determination(R2).Finally,the specific effects of the attention mechanism,convolutional layer,and fully-connected layer on the prediction performance of the model are systematically analyzed through ablation study.The results of experiment show that the GAN-LSTM-Attention model exhibits excellent performance and robustness in stock price prediction.

A Simulation Study to Assess Impacts of Habitat-Dependent Parameters of Surplus Production Model on Stock Assessment of Chilean Jack Mackerel in the Southeast Pacific Ocean

Pelagic fish are the most abundant species in upwelling regions,contributing 25%of total global fisheries production.Climate-driven changes in the marine environment play a crucial role in their population dynamics.Using Chilean jack mackerel(Trachurus murphyi)as an example,this study conducted simulations to quantify the impacts of environmental variations on the stock assessment.A habitat-based surplus production model was developed by integrating suitable habitat area into the model parameters carrying capacity(K)and intrinsic growth rate(r),with a suitable habitat area serving as the proxy for the environmental conditions for Chilean jack mackerel in the Southeast Pacific Ocean.The dynamics of Chilean jack mackerel stock and fisheries data were simulated,and four assessment models with different configurations were built to fit simulated data,with or without considering environmental effects.The results indicated that Joint K-r model,which integrated both parameters with the suitable habitat area index,outperformed the others by coming closest to the‘true'population dynamics.Ignoring habitat variations in the estimation model tended to overestimate biomass and underestimate harvest rate and reference points.Without observation and process error,the results were estimated with bias,while FMSY is relatively sensitive.This research illustrates the importance to consider random errors and environmental influences on populations,and provides foundation guidelines for future stock assessment.

Stock return prediction with multiple measures using neural network models

In the field of empirical asset pricing,the challenges of high dimensionality,non-linear relationships,and interaction effects have led to the increasing popularity of machine learning(ML)methods.This study investigates the performance of ML methods when predicting different measures of stock returns from various factor models and investigates the feature importance and interaction effects among firm-specific variables and macroeconomic factors in this context.Our findings reveal that neural network models exhibit consistent performance across different stock return measures when they rely solely on firm-specific characteristic variables.However,the inclusion of macroeconomic factors from the financial market,real economic activities,and investor sentiment leads to substantial improvements in the model performance.Notably,the degree of improvement varies with the specific measures of stock returns under consideration.Furthermore,our analysis indicates that,after the inclusion of macroeconomic factors,there is a dissimilarity in model performance,variable importance,and interaction effects among macroeconomic and firm-specific variables,particularly concerning abnormal returns derived from the Fama–French three-and five-factor models compared with excess returns.This divergence is primarily attributed to the extent to which these factor models remove the variance associated with the macroeconomic variables.These findings collectively offer valuable insights into the efficacy of neural network models for stock return predictions and contribute to a deeper understanding of the intricate relationship between factor models,stock returns,and macroeconomic conditions in the domain of empirical asset pricing.

Integrating remote sensing and 3-PG model to simulate the biomass and carbon stock of Larix olgensis plantation

Accurate estimations of biomass and its temporal dynamics are crucial for monitoring the carbon cycle in forest ecosystems and assessing forest carbon sequestration potentials.Recent studies have shown that integrating process-based models(PBMs)with remote sensing data can enhance simulations from stand to regional scales,significantly improving the ability to simulate forest growth and carbon stock dynamics.However,the utilization of PBMs for large-scale simulation of larch carbon storage distribution is still limited.In this study,we applied the parameterized 3-PG(Physiological Principles Predicting Growth)model across the Mengjiagang Forest Farm(MFF)to make broad-scale predictions of the biomass and carbon stocks of Larix olgensis plantation.The model was used to simulate average diameter at breast height(DBH)and total biomass,which were later validated with a wide range of observation data including sample plot data,forest management inventory data,and airborne laser scanning data.The results showed that the 3-PG model had relatively high accuracy for predicting both DBH and total biomass at stand and regional scale,with determination coefficients ranging from 0.78 to 0.88.Based on the estimation of total biomass,we successfully produced a carbon stock map of the Larix olgensis plantation in MFF with a spatial resolution of 20 m,which helps with relevant management advice.These findings indicate that the integration of 3-PG model and remote sensing data can well predict the biomass and carbon stock at regional and even larger scales.In addition,this integration facilitates the evaluation of forest carbon sequestration capacity and the development of forest management plans.

Tree mycorrhizal associations determine how biodiversity,large trees,and environmental factors drive aboveground carbon stock in temperate forests

Biodiversity,large trees,and environmental conditions such as climate and soil have important effects on forest carbon stocks.However,recent studies in temperate forests suggest that the relative importance of these factors depends on tree mycorrhizal associations,whereby large-tree effects may be driven by ectomycorrhizal(EM)trees,diversity effects may be driven by arbuscular mycorrhizal(AM)trees,and environment effects may depend on differential climate and soil preferences of AM and EM trees.To test this hypothesis,we used forest-inventory data consisting of over 80,000 trees from 631 temperate-forest plots(30 m×30 m)across Northeast China to examine how biodiversity(species diversity and ecological uniqueness),large trees(top 1%of tree diameters),and environmental factors(climate and soil nutrients)differently regulate aboveground carbon stocks of AM trees,EM trees,and AM and EM trees combined(i.e.total aboveground carbon stock).We found that large trees had a positive effect on both AM and EM tree carbon stocks.However,biodiversity and environmental factors had opposite effects on AM vs.EM tree carbon stocks.Specifically,the two components of biodiversity had positive effects on AM tree carbon stocks,but negative effects on EM tree carbon stocks.Environmental heterogeneity(mean annual temperature and soil nutrients)also exhibited contrasting effects on AM and EM tree carbon stocks.Consequently,for the total carbon stock,the positive large-tree effect far surpasses the diversity and environment effect.This is mainly because when integrating AM and EM tree carbon stock into total carbon stock,the opposite diversity-effect(also environment-effect)on AM vs.EM tree carbon stock counteracts each other while the consistent positive large-tree effect on AM and EM tree carbon stock is amplified.In summary,this study emphasized a mycorrhizal viewpoint to better understand the determinants of overarching aboveground carbon profile across regional forests.

Carbon stock estimation in halophytic wooded savannas of Uruguay:An ecosystem approach

Savannas constitute a mixture of trees and shrub patches with a more continuous herbaceous understory.The contribution of this biome to the soil organic carbon(SOC)and above-ground biomass(AGB)carbon(C)stock globally is significant.However,they are frequently subjected to land use changes,promoting increases in CO_(2) emissions.In Uruguay,subtropical wooded savannas cover around 100,000 ha,of which approximately 28%is circumscribed to sodic soils(i.e.,subtropical halophytic wooded savannas).Nevertheless,there is little background about the contribution of each ecosystem component to the C stock as well as site-specific allometric equations.The study was conducted in 5 ha of subtropical halophytic wooded savannas of the national protected area Esteros y Algarrobales del Rio Uruguay.This work aimed to estimate the contribution of the main ecosystem components(e.g.,soil,trees,shrubs,and herbaceous plants)to the C stock.Site-specific allometric equations for the most frequent tree species and shrub genus were fitted based on basal diameter(BD)and total height(H).The fitted equations accounted for between 77%and 98%of the aerial biomass variance of Netuma affinis and Vachellia caven.For shrubs(Baccharis sp.),the adjusted equation accounted for 86%of total aerial biomass.C stock for the entire system was 116.71±11.07 Mg·ha^(-1),of which 90.7%was allocated in the soil,8.3%in the trees,0.8%in the herbaceous plants,and 0.2%in the shrubs.These results highlight the importance of subtropical halophytic wooded savannas as C sinks and their relevance in the mitigation of global warming under a climate change scenario.

Stock Volatility Increases the Mortality Risk of Major Adverse Cardiovascular Events and Suicide:A Case-Crossover Study of 12 Million Deaths

Stock volatility constitutes an adverse psychological stressor,but few large-scale studies have focused on its impact on major adverse cardiovascular events(MACEs)and suicide.Here,we conducted an individual-level time-stratified case-crossover study to explore the association of daily stock volatility(daily returns and intra-daily oscillations for three kinds of stock indices)with MACEs and suicide among more than 12 million individual decedents from all counties in the mainland of China between 2013 and 2019.For daily stock returns,both stock increases and decreases were associated with increased mortal-ity risks of all MACEs and suicide.There were consistent and positive associations between intra-daily stock oscillations and mortality due to MACEs and suicide.The excess mortality risks occurred at the cur-rent day(lag 0 d),persisted for two days,and were greatest for suicide and hemorrhagic stroke.Taking the present-day Shanghai and Shenzhen 300 Index as an example,a 1%decrease in daily returns was associated with 0.74%-1.04%and 1.77%increases in mortality risks of MACEs and suicide,respectively;the corresponding risk increments were 0.57%-0.85%and 0.92%for a 1%increase in daily returns and 0.67%-0.77%and 1.09%for a 1%increase in intra-daily stock oscillations.The excess risks were more pro-nounced among individuals aged 65-74 years,males,and those with lower education levels.Our findings revealed considerable health risks linked to sociopsychological stressors,which are helpful for the gov-ernment and general public to mitigate the immediate cardiovascular and mental health risks associated with stock market volatility.

Transcriptional regulation of MdPIN7 by MdARF19 during gravityinduced formation of adventitious root GSA in self-rooted apple stock

Self-rooted apple stock is widely used for apple production.However,the shallowness of the adventitious roots in self-rooted apple stock leads to poor performance in the barren orchards of China.This is because of the considerable difference in the development of a gravitropic set-point angle(GSA)between self-rooted apple stock and seedling rootstock.Therefore,it is crucial to study the molecular mechanism of adventitious root GSA in self-rooted apple stock for breeding self-rooted and deep-rooted apple rootstock cultivars.An apple auxin response factor MdARF19 functioned to establish the adventitious root GSA of self-rooted apple stock in response to gravity and auxin signals.MdARF19 bound directly to the MdPIN7 promoter,activating its transcriptional expression and thus regulating the formation of the adventitious root GSA in 12-2 self-rooted apple stock.However,MdARF19 influenced the expression of auxin efflux carriers(MdPIN3 and MdPIN10)and the establishment of adventitious root GSA of self-rooted apple stock in response to gravity signals by direct activation of MdFLP.Our findings provide new information on the transcriptional regulation of MdPIN7 by auxin response factor MdARF19 in the regulation of the adventitious root GSA of self-rooted apple stock in response to gravity and auxin signals.

The changes in soil organic carbon stock and quality across a subalpine forest successional series

Soil organic carbon(SOC)affects the function of terrestrial ecosystem and plays a vital role in global carbon cycle.Yet,large uncertainty still existed regarding the changes in SOC stock and quality with forest succession.Here,the stock and quality of SOC at 1-m soil profile were investigated across a subalpine forest series,including shrub,deciduous broad-leaved forest,broadleaf-conifer mixed forest,middle-age coniferous forest and mature coniferous forest,which located at southeast of Tibetan Plateau.The results showed that SOC stock ranged from 9.8 to29.9 kg·m^(-2),and exhibited a hump-shaped response pattern across the forest successional series.The highest and lowest SOC stock was observed in the mixed forest and shrub forest,respectively.The SOC stock had no significant relationships with soil temperature and litter stock,but was positively correlated with wood debris stock.Meanwhile,the average percentages of polysaccharides,lignins,aromatics and aliphatics based on FTIR spectroscopy were 79.89%,0.94%,18.87%and 0.29%,respectively.Furthermore,the percentage of polysaccharides exhibited an increasing pattern across the forest successional series except for the sudden decreasing in the mixed forest,while the proportions of lignins,aromatics and aliphatics exhibited a decreasing pattern across the forest successional series except for the sudden increasing in the mixed forest.Consequently,the humification indices(HIs)were highest in the mixed forest compared to the other four successional stages,which means that the SOC quality in mixed forest was worse than other successional stages.In addition,the SOC stock,recalcitrant fractions and HIs decreased with increasing soil depth,while the polysaccharides exhibited an increasing pattern.These findings demonstrate that the mixed forest had higher SOC stock and worse SOC quality than other successional stages.The high proportion of SOC stock(66%at depth of 20-100 cm)and better SOC quality(lower HIs)indicate that deep soil have tremendous potential to store SOC and needs more attention under global chan ge.

Structured Multi-Head Attention Stock Index Prediction Method Based Adaptive Public Opinion Sentiment Vector

The present study examines the impact of short-term public opinion sentiment on the secondary market,with a focus on the potential for such sentiment to cause dramatic stock price fluctuations and increase investment risk.The quantification of investment sentiment indicators and the persistent analysis of their impact has been a complex and significant area of research.In this paper,a structured multi-head attention stock index prediction method based adaptive public opinion sentiment vector is proposed.The proposedmethod utilizes an innovative approach to transform numerous investor comments on social platforms over time into public opinion sentiment vectors expressing complex sentiments.It then analyzes the continuous impact of these vectors on the market through the use of aggregating techniques and public opinion data via a structured multi-head attention mechanism.The experimental results demonstrate that the public opinion sentiment vector can provide more comprehensive feedback on market sentiment than traditional sentiment polarity analysis.Furthermore,the multi-head attention mechanism is shown to improve prediction accuracy through attention convergence on each type of input information separately.Themean absolute percentage error(MAPE)of the proposedmethod is 0.463%,a reduction of 0.294% compared to the benchmark attention algorithm.Additionally,the market backtesting results indicate that the return was 24.560%,an improvement of 8.202% compared to the benchmark algorithm.These results suggest that themarket trading strategy based on thismethod has the potential to improve trading profits.

Stock trend prediction method coupled with multilevel indicators

To systematically incorporate multiple influencing factors,the coupled-state frequency memory(Co-SFM)network is proposed.This model integrates Copula estimation with neural networks,fusing multilevel data information,which is then fed into downstream learning modules.Co-SFM employs an upstream fusion module to incorporate multilevel data,thereby constructing a macro-plate-micro data structure.This configuration helps identify and integrate characteristics from different data levels,facilitating a deeper understanding of the internal links within the financial system.In the downstream model,Co-SFM uses a state-frequency memory network to mine hidden frequency information within stock prices,and the multifrequency patterns of sequential data are modeled.Empirical results show that Co-SFM s prediction accuracy for stock price trends is significantly better than that of other models.This is especially evident in multistep medium and long-term trend predictions,where integrating multilevel data results in notably improved accuracy.

Integrated Optimization of Timetable and Rolling Stock Circulation under Mixed Usages of Short-length and Full-length Services in Urban Rail Transit

The rapid growth of passenger flow in urban rail transit has led to great service pressures for metro companies in organizing train services to provide higher transportation capacities in order to satisfy passengers' travel demand, especially on those metro lines with insufficient rolling stock. In order to cope with high passenger flow service pressure, a mixed integer nonlinear programming(MINLP) model is proposed to optimize the line plan, timetable and rolling stock circulation simultaneously, to reduce the number of rolling stocks and increase the number of full-length services. A two-step algorithm strategy is proposed. In the first stage, the train timetable is optimized under the assumption that all the train services are the full-length services. In the second stage, the rolling stock plan is optimized based on the timetable optimized in the first stage. To ensure a feasible rolling stock circulation, certain full-length services are shortened to the short-length services due to the limited number of rolling stocks. Numerical experiments are performed based on the real-life data of Shanghai Metro Line 8. Results show that the proposed method can efficiently optimize the timetable and rolling stock circulation of the whole operation day. The optimized results are beneficial for both the service and the operational costs.

Design Strategies and Practice Paths for Improving Urban Quality in the Stock Era

In the middle and later stages of urbanization development,the growth of the real estate industry will stagnate,and urban renewal will become the mainstream.With the advancement of urban renewal,there are still problems in improving the quality of cities in the stock era and their design strategies.This paper analyzed the Linping Old City organic renewal project and the Xishui River ecological governance project in the stock era of urban quality improvement by sorting out the current development status,historical background,planning types,and design strategies of quality improvement in the stock era from the perspective of urban renewal,combining with project overview,main problems,design methods,and design content.Urban renewal is the leading direction for promoting urban development and construction on a global scale,and countries formulate different plans and practices based on their local characteristics.Urban renewal strategies should be diversified,and focus on livable environments,urban characteristics,etc.,while considering human factors,green innovation,etc.,in order to achieve smart community management and enhance the economic and social benefits brought by urban attractiveness.For successful cases such as the Linping Old City and Xishui River ecological governance project,corresponding urban quality improvement strategies and implementation plans should be formulated according to local conditions,with emphasis on social participation and people’s livelihood improvement.This study can help urban planning pay more attention to rational utilization and upgrading of existing urban resources,adapt to the current urban development situation,and promote sustainable urban development.

Growth, Population Parameters and Stock Status of Sarotherodon galilaeus in Samandeni Reservoir, Burkina Faso

Mango tilapia, Sarotherodon galilaeus is one of the most caught fish species in the Samandeni multi-species fishing sites of which, few data on its biology and exploitation are available. The study aimed to Assess the stock status of S. galilaeus. Sampling was conducted from March, 2021 to February 2022 based on commercial fish catches to analyze growth parameters, first sexual maturity size and harvest status of the stock. A total of 572 specimens including 297 females and 275 males were examined. The stock assessment was performed by using the Length based Bayesian method of Biomass (LBB) and that of growth by the ELEFAN method. The growth parameters showed a seasonality of growth and females appeared to grow faster than males. On the other hand, males had a greater asymptotic length than females. Results on the estimated length of fish at first maturity showed that females firstly reached the maturity compared to males. The relative biomass (B/B0) estimated for the stock was higher than the relative biomass that produces maximum sustainable yield (BMSY/B0) indicating healthy biomass. In addition, the length at first sexual maturity was less than the length at the first catch, indicating the absence of overfishing of growth. In addition, extending the study to the various stocks of the reservoir would be important for the sustainable management of the Samandeni high economic fishing area.

The Impact of Short Selling Disclosure Regulatory Constraint on the Lending Market and Stock Ownership

We examine the impact of the short sell disclosure(SSD)regime on the stock lending market and investor behaviors,employing a staggered difference-indifference(DiD)methodology.Our research reveals that the introduction of the disclosure regime enhances market transparency,resulting in a diminished appeal of stock ownership in the lending market for active investors.This shift is accompanied by a reduction in information leakage risks and longer loan durations.Specifically,our analysis reveals a significant decrease in the risk of loan recall by 4.87%,accompanied by an average increase of 23.72%in loan duration for short selling activities.Furthermore,the cost associated with short-sell disclosure causes a decline in both lending supply and short demand.

Exploring the Relationship Between Patent Forward Citation and Stock Return Rate Using Empirical Data of China Stock Market

A novel indicator called price-citation was proposed.Based on the company integrated patent database of China listed companies of common stocks(A-shares)with the stock price and the stock return rate data,more than two thousand of A-shares from 2017 to 2020 were selected.The effect of the traditional patent forward citation and the price-citation for discriminating the stock return rate was thoroughly analyzed via ANOVA.The A-shares of forward citation counts above the average showed higher stock return rate means than the A-shares having patents but receiving no forward citations.The price-citation,combining both the financial and patent attributes,defined as the multiplication of the current stock price and the currently receiving forward citation count,showed its excellence in discriminating the stock return rate.The A-shares of higher price-citation showed significantly higher stock return rate means while the A-shares of lower price-citation showed significantly lowest stock return rate means.The price-citation effect had not been changed by COVID-19 though COVID-19 affected the social and economic environment to a considerable extent in 2020.

Stock Type Prediction Based on Multiple Machine Learning Methods

Stocks in the Chinese stock market can be divided into ST stocks and normal stocks, so to prevent investors from buying potential ST stocks, this paper first performs SMOTEENN oversampling data preprocessing for the ST stock category, and selects 139 financial indicators and technical factor as predictive features. Then, it combines the Boruta algorithm and Copula entropy method for feature selection, effectively improving the machine learning model’s performance in ST stock classification, with the AUC values of the two models reaching 98% on the test set. In the model selection and optimization, this paper uses six major models, including logistic regression, XGBoost, AdaBoost, LightGBM, Catboost, and MLP, for modeling and optimizes them using the Optuna framework. Ultimately, XGBoost model is selected as the best model because its AUC value exceeds 95% and its running time is less. Finally, the XGBoost model is explained using the SHAP theory and the interaction between features is discovered, further improving the model’s accuracy and AUC value by about 0.6%, verifying the effectiveness of the model.

Stock Price Prediction and Traditional Models: An Approach to Achieve Short-, Medium- and Long-Term Goals

A comparative analysis of deep learning models and traditional statistical methods for stock price prediction uses data from the Nigerian stock exchange. Historical data, including daily prices and trading volumes, are employed to implement models such as Long Short Term Memory (LSTM) networks, Gated Recurrent Units (GRUs), Autoregressive Integrated Moving Average (ARIMA), and Autoregressive Moving Average (ARMA). These models are assessed over three-time horizons: short-term (1 year), medium-term (2.5 years), and long-term (5 years), with performance measured by Mean Squared Error (MSE) and Mean Absolute Error (MAE). The stability of the time series is tested using the Augmented Dickey-Fuller (ADF) test. Results reveal that deep learning models, particularly LSTM, outperform traditional methods by capturing complex, nonlinear patterns in the data, resulting in more accurate predictions. However, these models require greater computational resources and offer less interpretability than traditional approaches. The findings highlight the potential of deep learning for improving financial forecasting and investment strategies. Future research could incorporate external factors such as social media sentiment and economic indicators, refine model architectures, and explore real-time applications to enhance prediction accuracy and scalability.

Applying AI Models for Stock Investment Decisions

In the realm of finance and economics, the search for reliable stock predictions remains a focal point for researchers [1]. Contrary to common belief, stock market prices are not merely random guesses, but rather powerful tools that allow people to decide where to put their money, proving to be a significant aspect of finance. In this paper, by applying machine learning techniques, we propose to predict stock prices based on trends from previous years’ stock data using learning models, such as Linear Regression, MLP Regressor, Decision Tree Regressor and Random Forest Regressor. To enhance the model’s decision-making capabilities, the model was programmed to decide whether to sell or buy stocks using the predictions from the linear model. If the model anticipates an increase in stock prices, it suggests buying more stocks. On the contrary, if the model predicts a downturn, it suggests selling stocks in order to benefit the investor and enhance profitability. If the investor began with no stocks and $20,000, through the use of our model, the investor was able to make 161.3% profit. In another scenario where the investor holds 200 stocks and $10,000, the investor was able to make a 546.3% profit. Ultimately, the model results in profitable outcomes.

Comparative Analysis of Machine Learning Models for Stock Price Prediction: Leveraging LSTM for Real-Time Forecasting

The research focuses on improving predictive accuracy in the financial sector through the exploration of machine learning algorithms for stock price prediction. The research follows an organized process combining Agile Scrum and the Obtain, Scrub, Explore, Model, and iNterpret (OSEMN) methodology. Six machine learning models, namely Linear Forecast, Naive Forecast, Simple Moving Average with weekly window (SMA 5), Simple Moving Average with monthly window (SMA 20), Autoregressive Integrated Moving Average (ARIMA), and Long Short-Term Memory (LSTM), are compared and evaluated through Mean Absolute Error (MAE), with the LSTM model performing the best, showcasing its potential for practical financial applications. A Django web application “Predict It” is developed to implement the LSTM model. Ethical concerns related to predictive modeling in finance are addressed. Data quality, algorithm choice, feature engineering, and preprocessing techniques are emphasized for better model performance. The research acknowledges limitations and suggests future research directions, aiming to equip investors and financial professionals with reliable predictive models for dynamic markets.