A novel method for prediction of diabetes mellitus using deep convolutional neural network and long short-term memory

Kumari, Gorli L Aruna; Padmaja, Poosapati; Suma, Jaya G

doi:10.11591/ijeecs.v26.i1.pp404-413

Cited by 4 publications

(3 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Sailasya and Kumari (2021) analyzed the performance of stroke prediction using ML classification algorithms moreover, supervised backpropagation is applied to the training dataset to fine-tune the network. Kumari et al (2022) presented a novel method for the prediction of diabetes mellitus using deep CNN and long shortterm memory. Rajeswari et al (2019) designed a scheme that relies on Machine learning.…”

Section: Literature Reviewmentioning

confidence: 99%

Comparative study on early recognition and identifying diabetic retinopathy with different layers in CNN

Kumari¹,

Padmaja²,

Suma³

2022

Int. j. adv. appl. sci.

Self Cite

View full text Add to dashboard Cite

Diabetes is the most prevalent condition worldwide, and diabetic retinopathy (DR) is a subsequent condition caused by acute diabetic cases. It causes severe degeneration of the retina. The compounding blood vessels bloat and often burst, causing fluid leaks in the aqueous humor. This, in turn, causes the creation of undesirable nerve fiber infractions from the occlusion of arteries. Diagnosis requires a manual retinal examination that can often be inconsistent and deliberate with potential flaws in the diagnosis. Early detection through an ophthalmologist is paramount to prevent the prognosis of severe vision loss. Considering the current leap of machine learning in the field of healthcare, early detection of DR can be potentially made efficient with intelligent systems. This research proposes methodologies to fine-tune the existing pre-trained architectures, attaining the classification accuracies of 98% to classify the ocular fundus images which identify early prediction of diabetes. Additionally, this study presents an exposition of other equally scrutinized approaches to ultimately showcase a deep neural network architecture that can precisely classify normal fundus and degenerated fundus from the lowest to the most severe hierarchy. Among several layers in the CNN model pre-tuning and post-tuning exception layers outperformed with good results.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Comparative study on early recognition and identifying diabetic retinopathy with different layers in CNN

Kumari¹,

Padmaja²,

Suma³

2022

Int. j. adv. appl. sci.

Self Cite

View full text Add to dashboard Cite

show abstract

“…The ability of the long short-term memory (LSTM) network to learn order dependencies in sequence prediction problems in data series [31]- [36] makes it widely used for captioning tasks in generating sentence predictions. Research [37]- [40] utilizes a combination of CNN as feature extraction and LSTM to predict the output based on the order dependencies.…”

Section: Introductionmentioning

confidence: 99%

Image captioning to aid blind and visually impaired outdoor navigation

Faurina

Jelita

Vatresia

et al. 2023

IJ-AI

View full text Add to dashboard Cite

Artificial intelligence technology has dramatically improved the quality of services for human needs, one of which is technology to improve the quality of services for the blind and visually impaired, particularly technology that can help them understand visual sights to facilitate navigation in their daily lives. This study developed an image captioning model to aid the blind and visually impaired in outdoor navigation. The image captioning model employs the encoder-decoder method, with the convolutional neural network (CNN) feature extraction and attention layer as encoders and the long short-term memory (LSTM) as decoders. ResNet101 and ResNet152 are used in the encoder to extract image features. The results of the extraction and caption are forwarded to the attention layer and the LSTM network. The attention layer uses the Bahdanau attention mechanism. The accuracy of the model is calculated using the bilingual evaluation understudy score (BLEU), metric for evaluation of translation with explicit ordering (METEOR) and recall-oriented understudy for gisting evaluation-longest common subsequence (ROUGE-L). ResNet101 performed the best on BLEU-4, scoring 91.811% and 94.0337% in the METEOR evaluation. The captioning results show that the model is quite successful in displaying a simple caption that is suitable for each image.

show abstract

“…Recent studies have demonstrated that utilizing sensor data from a CGM device as the only input, machine learning models may be used to estimate future blood glucose levels and enhance the treatment of diabetic condition [22]- [27]. The last 20 minutes' worth of blood glucose values were used as the input by Pérez-Gandia et al [28] to predict the future blood glucose levels of 15 T1D patients.…”

Section: Introductionmentioning

confidence: 99%

Utilizing deep neural network for web-based blood glucose level prediction system

Alfian

Saputra

Subekti

et al. 2023

IJEECS

View full text Add to dashboard Cite

Machine learning algorithms can be used to forecast future blood glucose (BG) levels for diabetes patients, according to recent studies. In this study, dataset from continuous glucose monitoring (CGM) system was used as the sole input for the machine learning models. To forecast blood glucose levels 15, 30, and 45 minutes in the future, we suggested deep neural network (DNN) and tested it on 7 patients with type 1 diabetes (T1D). The suggested prediction model was evaluated against a variety of machine learning models, such as k-nearest neighbor (KNN), support vector regression (SVR), decision tree (DT), adaptive boosting (AdaBoost), random forest (RF), and eXtreme gradient boosting (XGBoost). The experimental findings demonstrated that the proposed DNN model outperformed all other models, with average root mean square errors (RMSEs) of 17.295, 25.940, and 35.146 mg/dL over prediction horizons (PHs) of 15, 30, and 45 minutes, respectively. Additionally, we have included the suggested prediction model in web-based blood glucose level prediction tools. By using this web-based system, patients may readily acquire their future blood glucose levels, allowing for the generation of preventative alarms prior to crucial hypoglycemia or hyperglycemic situations.

show abstract

A novel method for prediction of diabetes mellitus using deep convolutional neural network and long short-term memory

Cited by 4 publications

References 21 publications

Comparative study on early recognition and identifying diabetic retinopathy with different layers in CNN

Comparative study on early recognition and identifying diabetic retinopathy with different layers in CNN

Image captioning to aid blind and visually impaired outdoor navigation

Utilizing deep neural network for web-based blood glucose level prediction system

Contact Info

Product

Resources

About