A comprehensive review of explainable AI for disease diagnosis

Array, Год журнала: 2024, Номер 22, С. 100345 - 100345

Опубликована: Апрель 26, 2024

Nowadays, artificial intelligence (AI) has been utilized in several domains of the healthcare sector. Despite its effectiveness settings, massive adoption remains limited due to transparency issue, which is considered a significant obstacle. To achieve trust end users, it necessary explain AI models' output. Therefore, explainable (XAI) become apparent as potential solution by providing transparent explanations In this review paper, primary aim articles that are mainly related machine learning (ML) or deep (DL) based human disease diagnoses, and model's decision-making process explained XAI techniques. do that, two journal databases (Scopus IEEE Xplore Digital Library) were thoroughly searched using few predetermined relevant keywords. The PRISMA guidelines have followed determine papers for final analysis, where studies did not meet requirements eliminated. Finally, 90 Q1 selected in-depth covering Then, summarization findings presented, appropriate responses proposed research questions outlined. addition, challenges case diagnosis future directions sector presented.

Язык: Английский

---

An End-to-End framework for extracting observable cues of depression from diary recordings

Expert Systems with Applications, Год журнала: 2024, Номер 257, С. 125025 - 125025

Опубликована: Авг. 8, 2024

Because of the prevalence depression, its often-chronic course, relapse and associated disability, early detection non-intrusive monitoring is a crucial tool for timely diagnosis treatment, remission depression prevention relapse. In this way, impact on quality life well-being can be limited. Current attempts to use artificial intelligence classification are mostly data-driven thus non-transparent lack effective means deal with uncertainties. Therefore, in paper, we propose an end-to-end framework extracting observable cues from diary recordings. Furthermore, also explore feasibility automatic symptoms using behavioural cues. The proposed was used evaluate 28 video recordings Symptom Media dataset 27 DAIC-WOZ dataset. We compared presence extracted features between individuals without depressive disorder. identified several consistent previous studies terms their differentiation disorder across both datasets among language (i.e., negatively valanced words, first-person singular pronouns, some complexity, explicit mentions treatment depression), speech monotonous speech, voiced pauses, speaking rate, low articulation rate), facial rotational energy head movements). nature/context discourse, other disorders physical/psychological stress, resolution all play important role matching digital relevant background. work presented paper provides novel approach wide range opens up new opportunities further research.

Язык: Английский

---

Ensemble of hybrid model based technique for early detecting of depression based on SVM and neural networks

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Окт. 26, 2024

The prevalence of depression has increased dramatically over the last several decades: it is frequently overlooked and can have a significant impact on both physical mental health. Therefore, crucial to develop an automated detection system that instantly identify whether person depressed. Currently, machine learning (ML) artificial neural networks (ANNs) are among most promising approaches for developing computer-based systems predict health issues, such as depression. This study propose ensemble hybrid model-based techniques aims build strong model considers many psychological sociodemographic characteristics individual detect Support vector machines (SVM) multilayer perceptrons (MLP) two fundamental methods used construct suggested approach. DeprMVM served meta-learner. In this study, level-1 learner, whereas SVM MLP level-0 learners. After classifiers trained tested at level 0, their outputs based independent dependent variables in new data set was train meta-classifier. training class imbalance reduced by applying synthetic minority oversampling technique (SMOTE) cluster sampling together, which improved accuracy detecting Additionally, effectively reduce risk over-fitting from simply duplicating points. To further confirm effectiveness proposed method, various performance evaluation metrics were calculated compared with previous studies conducted specific dataset. conclusion, all identifying depression, approach had best accuracy, 99.39%, F1-score 99.51%.

Язык: Английский

---

Prediction of perinatal depression among women in Pakistan using Hybrid RNN-LSTM model

PeerJ Computer Science, Год журнала: 2025, Номер 11, С. e2673 - e2673

Опубликована: Фев. 26, 2025

Perinatal depression (PND) refers to a complex mental health condition that can occur during pregnancy (prenatal period) or in the first year after childbirth (postnatal period). Prediction of PND holds considerable importance due its significant role safeguarding and overall well-being both mothers their infants. Unfortunately, is difficult diagnose at an early stage thus may elevate risk suicide pregnancy. In addition, it contributes development postnatal depressive disorders. Despite gravity problem, resources for developing training AI models this area remain limited. To end, work, we have locally curated novel dataset named PERI DEP using Patient Health Questionnaire (PHQ-9), Edinburgh Postnatal Depression Scale (EPDS), socio-demographic questionnaires. The consists 14,008 records women who participated hospitals Lahore Gujranwala regions. We used SMOTE GAN oversampling data augmentation on set solve class imbalance problem. Furthermore, propose deep-learning framework combining recurrent neural networks (RNN) long short-term memory (LSTM) architectures. results indicate our hybrid RNN-LSTM model with achieves higher accuracy 95% F1 score 96%. Our study reveals prevalence rate among Pakistan (73.1%) indicating need prioritize prevention intervention strategies overcome public challenge.

Язык: Английский

---

Text Classification Experiments on Contextual Graphs Built by N-Gram Series

Communications in computer and information science, Год журнала: 2025, Номер unknown, С. 312 - 326

Опубликована: Янв. 1, 2025

Язык: Английский

---