Advanced Machine Learning for Financial Markets: A PCA-GRU-LSTM Approach

Journal of the Knowledge Economy, Journal Year: 2024, Volume and Issue: unknown

Published: June 1, 2024

Language: Английский

---

Machine learning models to predict nitrate concentration in a river basin

Environmental Research Communications, Journal Year: 2022, Volume and Issue: 4(12), P. 125012 - 125012

Published: Dec. 1, 2022

Abstract Aquifer-stream interactions affect the water quality in Mediterranean areas; therefore, coupling of surface and groundwater models is generally used to solve water-planning pollution problems river basins. However, their use limited because model inputs outputs are not spatially temporally linked, data update fitting laborious tasks. Machine learning have shown great potential simulation, as they can identify statistical relationship between input output without explicit requirement knowing physical processes. This allows ecological, hydrological, environmental variables that influence be analysed with a holistic approach. In this research, feature selection (FS) methods algorithms artificial intelligence—random forest (RF) eXtreme Gradient Boosting (XGBoost) trees—are simulate nitrate concentration determine main drivers related streams. The developed included 19 sampling 159 quality-gauging stations explanatory variables. were trained on 70 percent data, 30 validate predictions. Results showed combination FS method local knowledge about dataset best option improve model’s performance, while RF XGBoost high performance (r = 0.93 r 0.92, respectively). final ranking, based relative importance models, that, regarding nitrogen phosphorus concentration, location explained 87 variability. predicted accuracy using conditions or parameters entry enabled observation different relationships drivers. Thus, it possible delimit zones spatial risk approaches implementing solutions.

Language: Английский

---

Study of rail damage diagnosis and localization method based on intelligent wireless acoustic sensor network

Measurement, Journal Year: 2023, Volume and Issue: 217, P. 113032 - 113032

Published: May 17, 2023

Language: Английский

---

COVID-19 Media Chatter and Macroeconomic Reflectors on Black Swan: A Spanish and Indian Stock Markets Comparison

Risks, Journal Year: 2023, Volume and Issue: 11(5), P. 94 - 94

Published: May 16, 2023

Predictive analytics of financial markets in developed and emerging economies during the COVID-19 regime is undeniably challenging due to unavoidable uncertainty profound proliferation negative news on different platforms. Tracking media echo crucial explaining anticipating abrupt fluctuations markets. The present research attempts propound a robust framework capable channeling macroeconomic reflectors essential chatter-linked variables draw precise forecasts future figures for Spanish Indian stock predictive structure combines Isometric Mapping (ISOMAP), which non-linear feature transformation tool, Gradient Boosting Regression (GBR), an ensemble machine learning technique perform modelling. Explainable Artificial Intelligence (XAI) used interpret black-box type model infer meaningful insights. overall results duly justify incorporation local global chatter indices dynamics respective findings imply marginally better predictability than their counterparts. current work strives compare contrast reaction developing pandemic, has been argued share close resemblance Black Swan event when applying framework. insights linked dependence indicators can be leveraged policy formulations augmenting household finance.

Language: Английский

---

Understanding the determinants of vaccine hesitancy in the United States: A comparison of social surveys and social media

PLoS ONE, Journal Year: 2024, Volume and Issue: 19(6), P. e0301488 - e0301488

Published: June 6, 2024

The COVID-19 pandemic prompted governments worldwide to implement a range of containment measures, including mass gathering restrictions, social distancing, and school closures. Despite these efforts, vaccines continue be the safest most effective means combating such viruses. Yet, vaccine hesitancy persists, posing significant public health concern, particularly with emergence new variants. To effectively address this issue, timely data is crucial for understanding various factors contributing hesitancy. While previous research has largely relied on traditional surveys information, recent sources data, as media, have gained attention. However, potential media reliable proxy information population hesitancy, especially when compared survey remains underexplored. This paper aims bridge gap. Our approach uses social, demographic, economic predict levels in ten populous US metropolitan areas. We employ machine learning algorithms compare set baseline models that contain only variables incorporate separately. results show XGBoost algorithm consistently outperforms Random Forest Linear Regression, marginal differences between XGBoost. was case or thus highlighting promise latter complementary source. Results also reveal variations influential across five classes, age, ethnicity, occupation, political inclination. Further, application different MSAs yields mixed results, emphasizing uniqueness communities need approaches. In summary, study underscores data’s emphasizes importance tailoring interventions specific communities, suggests value combining sources.

Language: Английский

---

Evaluation of Growth Recovery Grade in Lodging Maize via UAV-based Hyperspectral Images

Journal of Remote Sensing, Journal Year: 2024, Volume and Issue: 4

Published: Jan. 1, 2024

Rapid and nondestructive monitoring of the temporal dynamic changes agronomic traits lodging maize is crucial for evaluating growth recovery status. The purpose this study to assess time-series in after using unmanned aerial vehicle (UAV) hyperspectral technology. Based on Entropy method, canopy height (CH) coverage (CC) were used represent structure index (CSI), while leaf chlorophyll content (LCC) plant water (PWC) physiological activity (PAI). theory normal (skewed) distribution, grade (GRG) was divided based estimated CSI PAI values. main results as follows: (a) With advance days (DAL), CH decreased increasing, other exhibited a downward trend. (b) R 2 values CH, CC, LCC, PWC estimation model 0.75, 0.69, 0.54, 0.49, respectively, MAPE 14.03%, 8.84%, 16.62%, 6.22%, testing set. (c) classified threshold PAI, achieving an overall accuracy 77.68%. Therefore, method proved effective damage. This provided reference efficient UAV-based images.

Language: Английский

---

Development and Validation of Machine Learning Models for Predicting Tumor Progression in OSCC

Oral Diseases, Journal Year: 2024, Volume and Issue: unknown

Published: Oct. 27, 2024

ABSTRACT Objectives Development of a prediction model using machine learning (ML) method for tumor progression in oral squamous cell carcinoma (OSCC) patients would provide risk estimation individual patient outcomes. Patients and Methods This predictive modeling study was conducted 1163 with OSCC from Hospital Stomatology, SYSU Cancer Center March 2009 to October 2021. Clinical, pathological, hematological features the were collected. Six ML algorithms explored, performance assessed by accuracy, sensitivity, specificity, f1 score, AUC. SHAP values used identify variables greatest contribution model. Results Among (mean [SD] age, 55.36 [12.91] years), 563 are development cohort 600 validation cohort. The Logistic Regression algorithm outperformed all other models, sensitivity 94.7% (68.2%), specificity 55.3% (63.7%), AUC 0.76 ± 0.09 (0.723) (validation) most feature neutrophil count. Conclusion demonstrated models can improve clinical through basic information patients. These tools could be may help direct intervention.

Language: Английский

---

Can financial stress be anticipated and explained? Uncovering the hidden pattern using EEMD-LSTM, EEMD-prophet, and XAI methodologies

Complex & Intelligent Systems, Journal Year: 2022, Volume and Issue: 9(4), P. 4169 - 4193

Published: Dec. 26, 2022

Global financial stress is a critical variable that reflects the ongoing state of several key macroeconomic indicators and markets. Predictive analytics stress, nevertheless, has seen very little focus in literature as now. Futuristic movements markets can be anticipated if same predicted with satisfactory level precision. The current research resorts to two granular hybrid predictive frameworks discover inherent pattern across variables geography. structure utilizes Ensemble Empirical Mode Decomposition (EEMD) for time series decomposition. Long Short-Term Memory Network (LSTM) Facebook's Prophet algorithms are invoked on top decomposed components scrupulously investigate predictability final regulated by Office Financial Research (OFR). A rigorous feature screening using Boruta methodology been utilized too. findings exercises reveal assets continents accurately short long-run horizons even at steep distress during COVID-19 pandemic. appear statistically significant expense model interpretation. To resolve issue, dedicated Explainable Artificial Intelligence (XAI) methods have used interpret same. immediate past information largely explains patterns long run, while short-run fluctuations tracked closely monitoring technical indicators.

Language: Английский

---

Modelización integrada con aprendizaje automático para evaluar la contaminación por nutrientes en las masas de agua actual y bajo el efecto del cambio climático. Aplicación a la Demarcación Hidrográfica del Júcar

Published: Jan. 24, 2024

A Fabiola Fandiño por cuidar a

---

Harnessing the Power of Prophet Algorithm for Advanced Predictive Modeling of Grab Holdings Stock Prices

Journal of Applied Data Sciences, Journal Year: 2024, Volume and Issue: 5(2), P. 326 - 341

Published: May 15, 2024

This study investigates the effectiveness of Prophet algorithm in predicting Grab Holdings' stock prices dataset from Kaggle. By meticulously analyzing historical closing, high, low, and volume data, research aims to uncover market patterns gain insights into investor sentiment based on short-term forecasting. The findings reveal a dynamic trajectory for stock, characterized by significant fluctuations evolving confidence. reached peak $14 early 2022, indicating optimism, but subsequently experienced decline $4 late 2023, reflecting shift sentiment. Notably, 2023 witnessed heightened volatility compared evident more price swings increased trading volume. demonstrated promising potential prediction better than traditional methods, which overlook presence seasonality or fail adapt conditions, leading less accurate forecasts. excellent performance is indicated Mean Absolute Percentage Error (MAPE) 10.45511%, (MAE) 3.112026, Root Squared (RMSE) 3.516969. Compared ARIMA, MAE RMSE resulting are much lower their counterparts, 14.49675 16.079898, respectively. These widely used metrics suggest moderate accuracy future prices. offers valuable investors that they can use understand trend make informed investment decisions regarding buying selling opportunities. However, it crucial acknowledge inherent limitations such models interpret results cautiously, considering ever-changing dynamics financial market.

Language: Английский

---