Communications in computer and information science, Journal Year: 2025, Volume and Issue: unknown, P. 312 - 326
Published: Jan. 1, 2025
Language: Английский
Communications in computer and information science, Journal Year: 2025, Volume and Issue: unknown, P. 312 - 326
Published: Jan. 1, 2025
Language: Английский
Computer Methods and Programs in Biomedicine, Journal Year: 2022, Volume and Issue: 226, P. 107161 - 107161
Published: Sept. 27, 2022
Language: Английский
Citations
444Electronics, Journal Year: 2022, Volume and Issue: 11(5), P. 676 - 676
Published: Feb. 23, 2022
Depression is a prevalent sickness, spreading worldwide with potentially serious implications. Timely recognition of emotional responses plays pivotal function at present, the profound expansion social media and users internet. Mental illnesses are highly hazardous, stirring more than three hundred million people. Moreover, that why research focused on this subject. With advancements machine learning availability sample data relevant to depression, there possibility developing an early depression diagnostic system, which key lessening number afflicted individuals. This paper proposes productive model by implementing Long-Short Term Memory (LSTM) model, consisting two hidden layers large bias Recurrent Neural Network (RNN) dense layers, predict from text, can be beneficial in protecting individuals mental disorders suicidal affairs. We train RNN textual identify semantics, written content. The proposed framework achieves 99.0% accuracy, higher its counterpart, frequency-based deep models, whereas false positive rate reduced. also compare other models regarding mean accuracy. approach indicates feasibility LSTM achieving exceptional results for emotions numerous subscribers.
Language: Английский
Citations
146Computers in Biology and Medicine, Journal Year: 2023, Volume and Issue: 166, P. 107555 - 107555
Published: Oct. 4, 2023
In domains such as medical and healthcare, the interpretability explainability of machine learning artificial intelligence systems are crucial for building trust in their results. Errors caused by these systems, incorrect diagnoses or treatments, can have severe even life-threatening consequences patients. To address this issue, Explainable Artificial Intelligence (XAI) has emerged a popular area research, focused on understanding black-box nature complex hard-to-interpret models. While humans increase accuracy models through technical expertise, how actually function during training be difficult impossible. XAI algorithms Local Interpretable Model-Agnostic Explanations (LIME) SHapley Additive exPlanations (SHAP) provide explanations models, improving predictions providing feature importance increasing confidence systems. Many articles been published that propose solutions to problems using alongside explainability. our study, we identified 454 from 2018-2022 analyzed 93 them explore use techniques domain.
Language: Английский
Citations
124IEEE/CAA Journal of Automatica Sinica, Journal Year: 2023, Volume and Issue: 10(4), P. 859 - 876
Published: March 28, 2023
Artificial intelligence (AI) continues to transform data analysis in many domains. Progress each domain is driven by a growing body of annotated data, increased computational resources, and technological innovations. In medicine, the sensitivity complexity tasks, potentially high stakes, requirement accountability give rise particular set challenges. this review, we focus on three key methodological approaches that address some challenges AI-driven medical decision making. 1) Explainable AI aims produce human-interpretable justification for output. Such models increase confidence if results appear plausible match clinicians expectations. However, absence explanation does not imply an inaccurate model. Especially highly non-linear, complex are tuned maximize accuracy, such interpretable representations only reflect small portion justification. 2) Domain adaptation transfer learning enable be trained applied across multiple For example, classification task based images acquired different acquisition hardware. 3) Federated enables large-scale without exposing sensitive personal health information. Unlike centralized learning, where machine has access entire training federated process iteratively updates sites exchanging parameter updates, data. This narrative review covers basic concepts, highlights relevant corner-stone state-of-the-art research field, discusses perspectives.
Language: Английский
Citations
48ACM Transactions on Asian and Low-Resource Language Information Processing, Journal Year: 2022, Volume and Issue: 23(1), P. 1 - 20
Published: Nov. 5, 2022
Depression is a kind of emotion that negatively impacts people's daily lives. The number people suffering from long-term feelings increasing every year across the globe. Depressed patients may engage in self-harm behaviors, which occasionally result suicide. Many psychiatrists struggle to identify presence mental illness or negative early provide better course treatment before they reach critical stage. One most challenging problems detecting depression at earliest possible Researchers are using Natural Language Processing (NLP) techniques analyze text content uploaded on social media, helps design approaches for depression. This work analyses numerous prior studies used learning existing methods suffer model representation detect with high accuracy. present addresses solution these by creating new hybrid deep neural network representations called “Fasttext Convolution Neural Network Long Short-Term Memory (FCL).” In addition, this utilizes advantage NLP simplify analysis during development. FCL comprises fasttext embedding considering out-of-vocabulary (OOV) semantic information, convolution (CNN) architecture extract global and (LSTM) local features dependencies. was implemented real-world datasets utilized literature. proposed technique provides results than state-of-the-art
Language: Английский
Citations
42Beni-Suef University Journal of Basic and Applied Sciences, Journal Year: 2025, Volume and Issue: 14(1)
Published: Jan. 24, 2025
Abstract Background One of the psychological problems that have become very prevalent in modern world is depression, where mental health disorders common. Depression, as reported by WHO, second-largest factor worldwide burden illnesses. As these issues grow, social media has a tremendous platform for people to express themselves. A user’s behavior may therefore disclose lot about their emotional state and health. This research offers novel framework depression detection from Arabic textual data utilizing deep learning (DL), natural language processing (NLP), machine (ML), BERT transformers techniques light disease’s high prevalence. To do this, dataset tweets was used, which collected 3 sources, we mention later. The constructed two variants, one with binary classification other multi-classification. Results In classifications, used ML such “support vector (SVM), random forest (RF), logistic regression (LR), Gaussian naive Bayes (GNB),” “ARABERT.” comparison transformers, ARABERT accuracy 93.03 percent rate. multi-classification, DL “long short-term memory (LSTM),” “Multilingual BERT.” multilingual multi-classification an 97.8%. Conclusion Through user-generated content, can detect depressed using artificial intelligence technology fast manner instead medical technology.
Language: Английский
Citations
2Sensors, Journal Year: 2022, Volume and Issue: 22(24), P. 9775 - 9775
Published: Dec. 13, 2022
In today's world, mental health diseases have become highly prevalent, and depression is one of the problems that has widespread. According to WHO reports, second-leading cause global burden diseases. proliferation such issues, social media proven be a great platform for people express themselves. Thus, user's can speak deal about his/her emotional state health. Considering high pervasiveness disease, this paper presents novel framework detection from textual data, employing Natural Language Processing deep learning techniques. For purpose, dataset consisting tweets was created, which were then manually annotated by domain experts capture implicit explicit context. Two variations on having binary ternary labels, respectively. Ultimately, deep-learning-based hybrid Sequence, Semantic, Context Learning (SSCL) classification with self-attention mechanism proposed utilizes GloVe (pre-trained word embeddings) feature extraction; LSTM CNN used sequence semantics tweets; finally, GRUs used, focus contextual information in tweets. The outperformed existing techniques detecting context, an accuracy 97.4 labeled data 82.9 data. We further tested our SSCL unseen (random tweets), F1-score 94.4 achieved. Furthermore, order showcase strengths framework, we validated it "News Headline Data set" sarcasm detection, considering different domain. It also outmatched performance cross-domain validation.
Language: Английский
Citations
36Archives of Computational Methods in Engineering, Journal Year: 2023, Volume and Issue: 30(8), P. 4733 - 4756
Published: June 24, 2023
Language: Английский
Citations
18Archives of Computational Methods in Engineering, Journal Year: 2024, Volume and Issue: unknown
Published: April 16, 2024
Language: Английский
Citations
7Published: June 11, 2024
With increasing electronic medical data and the development of artificial intelligence, Clinical Decision Support Systems (CDSSs) assist clinicians in diagnosis prescription. Traditional knowledge-based CDSSs follow an accumulated knowledgebase a predefined rule system, which clarifies decision-making process; however, maintenance cost issues exist quality control standardization process. Non-knowledge-based utilize vast amounts algorithms to effectively decide; deep learning black-box problem causes unreliable results. EXplainable Artificial Intelligence (XAI)-based CDSS provides valid rationale explainable It ensures trustworthiness transparency by showing recommendation prediction results process through techniques. However, existing systems have limitations, such as scope utilization lack explanatory power AI models. This study proposes new XAI-based framework address these issues; introduce resources, datasets, models that can be utilized; foundation model support various disease domains. Finally, we propose future directions for technology highlight societal need addressed emphasize potential future.
Language: Английский
Citations
7