An Accurate Multiple Data Based Stock Prediction and Sentiment Analysis Using Synergic Deep Info Convolutional Neural Network

Computational Economics, Journal Year: 2025, Volume and Issue: unknown

Published: April 21, 2025

Language: Английский

---

Development of predictive model for predicting postoperative BMI and optimize bariatric surgery: a single center pilot study

Surgery for Obesity and Related Diseases, Journal Year: 2024, Volume and Issue: 20(12), P. 1234 - 1243

Published: July 8, 2024

BackgroundThe pilot study addresses the challenge of predicting postoperative outcomes, particularly body mass index (BMI) trajectories, following bariatric surgery. The complexity this task makes preoperative personalized obesity treatment challenging.ObjectivesTo develop and validate sophisticated machine learning (ML) algorithms capable accurately forecasting BMI reductions up to 5 years surgery aiming enhance planning care. secondary goal involves creation an accessible web-based calculator for healthcare professionals. This is first article that compares these methods in prediction.SettingThe was carried out from January 2012 December 2021 at GZOAdipositas Surgery Center, Switzerland. Preoperatively, data 1004 patients were available. Six months postoperatively, 1098 For time points 12 months, 18 2 years, 3 4 number follow-ups available: 971, 898, 829, 693, 589, 453.MethodsWe conducted a comprehensive retrospective review adult who underwent (Roux-en-Y gastric bypass or sleeve gastrectomy), focusing on individuals with data. Patients certain conditions those lacking complete sets excluded. Additional exclusion criteria incomplete follow-up, pregnancy during follow-up period, ≤30 kg/m2.ResultsThis analyzed 1104 patients, 883 used model training 221 final evaluation, achieved reliable predictive capabilities, as measured by root mean square error (RMSE). RMSE values three tasks 2.17 (predicting next value), 1.71 any future point), 3.49 5-year curve). These results showcased through web application, enhancing clinical accessibility decision-making.ConclusionThis highlights potential ML significantly improve surgical outcomes overall efficiency precise predictions intervention strategies.

Language: Английский

---

Light Recurrent Unit: Towards an Interpretable Recurrent Neural Network for Modeling Long-Range Dependency

Electronics, Journal Year: 2024, Volume and Issue: 13(16), P. 3204 - 3204

Published: Aug. 13, 2024

Recurrent neural networks (RNNs) play a pivotal role in natural language processing and computer vision. Long short-term memory (LSTM), as one of the most representative RNNs, is built upon relatively complex architecture with an excessive number parameters, which results large storage, high training cost, lousy interpretability. In this paper, we propose lightweight network called Light Unit (LRU). On hand, designed accessible gate structure, has interpretability addresses issue gradient disappearance. other introduce Stack Cell (SRC) structure to modify activation function, not only expedites convergence rates but also enhances network. Experimental show that our proposed LRU advantages fewer strong interpretability, effective modeling ability for variable length sequences on several datasets. Consequently, could be promising alternative traditional RNN models real-time applications space or time constraints, potentially reducing storage costs while maintaining performance.

Language: Английский

---

A deep fusion model for stock market prediction with news headlines and time series data

Neural Computing and Applications, Journal Year: 2024, Volume and Issue: 36(34), P. 21229 - 21271

Published: Aug. 24, 2024

Abstract Time series forecasting models are essential decision support tools in real-world domains. Stock market is a remarkably complex domain, due to its quickly evolving temporal nature, as well the multiple factors having an impact on stock prices. To date, number of machine learning-based approaches have been proposed literature tackle trend prediction. However, they typically tend analyze single data source or modality, consider modalities isolation and rely simple combination strategies, with potential reduction their modeling power. In this paper, we propose multimodal deep fusion model predict trends, leveraging daily prices, technical indicators, sentiment news headlines published by media outlets. The architecture leverages BERT-based branch fine-tuned financial long short-term memory (LSTM) that captures relevant patterns multivariate data, including prices indicators. Our experiments 12 different datasets demonstrate our more effective than popular baseline approaches, both terms accuracy trading performance portfolio analysis simulation, highlighting positive learning for

Language: Английский

---

Neural Network-Based Predictive Models for Stock Market Index Forecasting

Journal of risk and financial management, Journal Year: 2024, Volume and Issue: 17(6), P. 242 - 242

Published: June 11, 2024

The stock market, characterised by its complexity and dynamic nature, presents significant challenges for predictive analytics. This research compares the effectiveness of neural network models in predicting S&P500 index, recognising that a critical component financial decision making is market volatility. examines such as Long Short-Term Memory (LSTM), Convolutional Neural Network (CNN), Artificial (ANN), Recurrent (RNN), Gated Unit (GRU), taking into account their individual characteristics pattern recognition, sequential data processing, handling nonlinear relationships. These are analysed using key performance indicators Root Mean Square Error (RMSE), Absolute Percentage (MAPE), Directional Accuracy, metric considered essential prediction both training testing phases this research. results show although each model has own advantages, GRU CNN perform particularly well according to these metrics. lowest error metrics, indicating robustness accurate prediction, while highest directional accuracy testing, efficiency processing. study highlights potential combining metrics consideration when decisions due changing dynamics market.

Language: Английский

---

A Pork Price Prediction Model Based on a Combined Sparrow Search Algorithm and Classification and Regression Trees Model

Applied Sciences, Journal Year: 2023, Volume and Issue: 13(23), P. 12697 - 12697

Published: Nov. 27, 2023

The frequent fluctuation of pork prices has seriously affected the sustainable development industry. accurate prediction can not only help practitioners make scientific decisions but also them to avoid market risks, which is way promote healthy Therefore, improve accuracy prices, this paper first combines Sparrow Search Algorithm (SSA) and traditional machine learning model, Classification Regression Trees (CART), establish an SSA-CART optimization model for predicting prices. Secondly, based on Sichuan price data during 12th Five-Year Plan period, linear correlation between piglet, corn, fattening pig feed, was measured using Pearson coefficient. Thirdly, MAE fitness value calculated by combining validation set training set, hyperparameter “MinLeafSize” optimized via SSA. Finally, a comparative analysis performance White Shark Optimizer (WSO)-CART CART Simulated Annealing (SA)-CART demonstrated that best (compared with single decision tree, R2 increased 9.236%), conducive providing support prediction. great practical significance stabilizing production, ensuring growth farmers’ income, promoting sound economic development.

Language: Английский

---

A CNN-LSTM deep neural network with technical indicators and sentiment analysis for stock price forecastings

Published: Feb. 21, 2024

Language: Английский

---

A Multi-Aspect Informed GRU: A Hybrid Model of Flight Fare Forecasting with Sentiment Analysis

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(10), P. 4221 - 4221

Published: May 16, 2024

This paper presents an advanced method for forecasting flight fares that combines aspect-based sentiment analysis (ABSA) with deep learning techniques, particularly the gated recurrent unit (GRU) model. approach leverages historical airline ticket transaction data and customer reviews to better understand fare dynamics impact of sentiments on pricing. The aspect extracts key service aspects from feedback provides insightful correlations airfare. These were further categorized into nine groups sensitivity analysis, which offered a deeper understanding how each group influences customers’ attitudes. ABSA-driven marks departure traditional models by utilizing alongside improve predictive accuracy. Its effectiveness is demonstrated through metrics including root mean square error (RMSE) 0.0071, absolute (MAE) 0.0137, coefficient determination (R2) 0.9899. Additionally, this model shows strong prediction performance in both short- long-term predictions. It not only advances airfare methods but valuable insights decision makers industry refine pricing strategies or make improvements when it indicated some services require attention.

Language: Английский

---

Machine Learning Models-Based Forecasting Moroccan Stock Market

Lecture notes in networks and systems, Journal Year: 2024, Volume and Issue: unknown, P. 56 - 66

Published: Jan. 1, 2024

Language: Английский

---

Analyzing the critical steps in deep learning-based stock forecasting: a literature review

PeerJ Computer Science, Journal Year: 2024, Volume and Issue: 10, P. e2312 - e2312

Published: Sept. 23, 2024

Stock market or individual stock forecasting poses a significant challenge due to the influence of uncertainty and dynamic conditions in financial markets. Traditional methods, such as fundamental technical analysis, have been limited coping with uncertainty. In recent years, this has led growing interest using deep learning-based models for prediction. However, accuracy reliability these depend on correctly implementing series critical steps. These steps include data collection feature extraction selection, noise elimination, model selection architecture determination, choice training-test approach, performance evaluation. This study systematically examined literature, investigating effects model’s performance. review focused studies between 2020–2024, identifying influential by conducting systematic literature search across three different databases. The identified regarding seven essential creating successful reliable prediction were thoroughly examined. findings from examinations summarized tables, gaps detailed. not only provides comprehensive understanding current but also serves guide future research.

Language: Английский

---