Sentiment Informed Sentence BERT-Ensemble Algorithm for Depression Detection

Опубликована: Июль 16, 2024

The world health organisation (WHO) revealed approximately 280 million people in the suffer from depression. Yet, existing studies on early-stage depression detection using machine learning (ML) techniques are limited. Prior have applied a single stand-alone algorithm which unable to deal with data complexities, prone overfitting and limited generalisation. To this end, our paper examined performance of several ML algorithms for two benchmark social media datasets (D1 D2). More specifically, we incorporated sentiment indicator improve model performance. Our experimental results showed that sentence bidirectional encoder representations transformers (SBERT) numerical vectors fitted into stacking ensemble achieved comparable F1 scores 69% dataset (D1) 76% (D2). findings suggest utilising indicators as additional feature yields an improved thus, recommend development depressive term corpus future work.

Язык: Английский

---

A novel hybrid CNN-KNN ensemble voting classifier for Parkinson’s disease prediction from hand sketching images

Multimedia Tools and Applications, Год журнала: 2024, Номер unknown

Опубликована: Май 14, 2024

Язык: Английский

---

Explainable Multi-Layer Dynamic Ensemble Framework Optimized for Depression Detection and Severity Assessment

Diagnostics, Год журнала: 2024, Номер 14(21), С. 2385 - 2385

Опубликована: Окт. 25, 2024

Depression is a pervasive mental health condition, particularly affecting older adults, where early detection and intervention are essential to mitigate its impact. This study presents an explainable multi-layer dynamic ensemble framework designed detect depression assess severity, aiming improve diagnostic precision provide insights into contributing factors.

Язык: Английский

---

A Multifaceted Approach to Understanding Mental Health Crises in the COVID-19 Era

Advances in medical diagnosis, treatment, and care (AMDTC) book series, Год журнала: 2024, Номер unknown, С. 97 - 119

Опубликована: Янв. 5, 2024

The COVID-19 pandemic, starting in Wuhan, China December 2019, led to widespread health and economic challenges, causing millions of deaths globally. Beyond physical health, it triggered a mental crisis, especially during lockdowns. To understand address this, study collected data using 90 features the lockdown period. Machine learning (ML) was employed detect key impacting crises. Three ML algorithms—random forest, random tree, multilayer perceptron—were chosen. Random known for robustness, achieved 97.58% accuracy. supervised algorithm with decision trees, yielded 93.24% Multilayer perceptron (MLP), an artificial neural network, 94.20% accuracy by nonlinear relationships. A 10-fold cross-validation method used evaluate these models, enhancing performance reducing bias overfitting. It involves dividing into ten subsets, training on nine, evaluating remaining, repeating this times estimate true unseen data.

Язык: Английский

---

Sentiment Informed Sentence BERT-Ensemble Algorithm for Depression Detection

Big Data and Cognitive Computing, Год журнала: 2024, Номер 8(9), С. 112 - 112

Опубликована: Сен. 5, 2024

The World Health Organisation (WHO) revealed approximately 280 million people in the world suffer from depression. Yet, existing studies on early-stage depression detection using machine learning (ML) techniques are limited. Prior have applied a single stand-alone algorithm, which is unable to deal with data complexities, prone overfitting, and limited generalization. To this end, our paper examined performance of several ML algorithms for two benchmark social media datasets (D1 D2). More specifically, we incorporated sentiment indicators improve model performance. Our experimental results showed that sentence bidirectional encoder representations transformers (SBERT) numerical vectors fitted into stacking ensemble achieved comparable F1 scores 69% dataset (D1) 76% (D2). findings suggest utilizing as an additional feature yields improved performance, thus, recommend development depressive term corpus future work.

Язык: Английский

---

Predicting Depression and Anxiety: A Multi-Layer Perceptron for Analyzing the Mental Health Impact of COVID-19

Опубликована: Май 17, 2024

Язык: Английский

---

Early Risk Prediction of Depression Based on Social Media Posts in Arabic

Опубликована: Ноя. 6, 2023

Depression is a prevalent global health issue, impacting various aspects of individuals' lives, including home and social interactions. In the Arabic environment, stigma surrounding mental disorders limited awareness in psychiatry domain has made early diagnosis depression challenging task. However, media platforms have enabled individuals to express their thoughts personal experiences, making these valuable resource for monitoring. this paper, we propose an approach predict signs utilizing posts expressed on Twitter platform. The proposed methodology integrates knowledge extracted using LLM-based transformer, UMLS medical resource, machine learning prediction algorithms. To best our knowledge, first research study that maps translated texts external resources improve accuracy model. model consists four phases. Firstly, NLP-based data preprocessing pipeline employed ensure input dataset suitable format analysis. Secondly, ChatGPT transformer utilized translate tweets into English, enabling further processing analysis English. Thirdly, relevant concepts are from text quickUMLS tool metathesaurus, aiding identifying important terms related health. Fourthly, TF-IDF Bag Words (BOW) algorithms used assign weights features, highlighting significance concepts. Finally, classification algorithms, Support Vector Machine (SVM), Logistic Regression (LR), Random Forest (RF), Naive Bayes (NB), Stochastic Gradient Descent (SGD), trained Among classifiers, with demonstrated performance, achieving 80.24%.

Язык: Английский

---

A Review of Mental Health Analysis Through Social Media Using Machine Learning and Deep Learning Approaches

Опубликована: Май 23, 2024

Язык: Английский

---

Sentiment Analysis Based on Machine Learning Techniques: A Comprehensive Review

Indonesian Journal of Computer Science, Год журнала: 2024, Номер 13(3)

Опубликована: Июнь 15, 2024

In the landscape of digital communication, sentiment analysis stands out as a pivotal technology for deciphering vast troves unstructured text generated online. When integrated with machine learning, transforms into powerful tool capable distilling insights from complex human emotions and opinions expressed across social media, reviews, forums. This review paper embarks on thorough exploration integration learning techniques analysis, shedding light latest advancements, challenges, applications spanning various sectors including public health, finance, consumer behavior. It meticulously examines role in elevating through improved accuracy, adaptability, depth analysis. Furthermore, discusses implications these technologies understanding sentiment, tracking health trends, forecasting market movements. By synthesizing findings seminal studies cutting-edge research, this not only charts current but also forecasts trajectory underscores necessity ongoing innovation models to keep pace evolving discourse. The presented herein aim guide future research endeavors, highlight transformative impact outline potential new that could benefit society at large.

Язык: Английский

---

Risk Management in Hospitality: Machine Learningbased Approaches for Crisis Prediction and Response

Опубликована: Сен. 18, 2024

Язык: Английский

---