17 A REVIEW STUDY OF DEEP LEARNING ALGORITHMS AS A COMPUTER AID IN
HEALTHCARE SECTOR
1Marie Fernandes, 2Dr. Rekha Ranawat
Department of Computer Science, Renaissance University, Indore, India
Abstract: Deep learning is a broader part or a branch of machine learning. Deep learning are methods based on artificial neural networks (ANN) with representation learning that can be used as computer aid in solving complex problems. Deep learning method forms a foundation for predictive modeling. Learning can be supervised, semi-supervised or unsupervised. Artificial Neural Networks (ANN) is utilized in deep learning to research data and make predictions. In the field of healthcare, deep learning has made it possible to predict and diagnose the diseases through implementation of algorithms that are hierarchy of complexity and abstractions. As compared to traditional machine learning, the algorithms of traditional machine learning are linear and the learning process is supervised but in deep learning, the algorithms are stack of hierarchy and the learning can be both supervised as well as unsupervised.
The field of deep learning has witnessed sharp and striking advances in the ability of machines to understand and manipulate data, including images, languages and speech.
Healthcare and medicine sectors are benefitting immensely from deep learning just because of the volume of data being generated which is continuously growing as per the digital record systems. In this paper, a comprehensive review is presented on deep learning algorithms and the way it's utilized in healthcare to predict and diagnose various diseases.
Keywords: Deep learning, Healthcare, Algorithms, Disease.
I. INTRODUCTION
Machine learning as a part and application of Artificial Intelligence consists of various algorithms that analyze and parse the data, they learn from that data and the learnt data is then applied to make decisions that are well informed and well defined. On the other hand deep learning is considered as a subset and part of Machine Learning (ML). Being an Artificial Intelligence (AI) technology, it is capable of learning and developing without being explicitly programmed. Deep learning structures algorithms in layers to create an artificial neural network that can learn and make intelligent decisions on its own. Machine learning works on simple concepts while deep learning deals with Artificial Neural Networks (ANN), which are programmed to analyze, learn and respond to complex situations faster and quicker than humans. Deep learning works as an aid with speech recognition, image classification and language translation. Without human intervention it also can be used to solve any kind of pattern recognition problem.
ANN consists of many layers where each and every layer can perform complex operations such as representation and abstraction that make sense of text, images and sound.
As the human brain is formed from neurons, the neural network consists of layers of nodes. The nodes in each layer are linked to the nodes of the adjacent layers. As per the amount of layers in it, the network is claimed to be deeper. The neuron within the human brain receives thousands of signals from other neurons and in the same manner in an ANN, signal travel between nodes and weights are assigned to the corresponding nodes. A node which has a heavy weight will provide more effect on the nodes within the next layer. The output is produced after compiling the weighted input.
Deep learning algorithms are capable of analyzing vast volumes of data and facts and thus they provide reliable information. Artificial neural networks interpret data based on the answers to a sequence of binary true or false questions requiring extremely abstract mathematical equations when analyzing the input file. A face recognition technology, for instance, learns to spot and distinguish edges and features of faces, then trains to recognize more important portions of the faces, and eventually faces as an entire. Similarly, the software trains itself at each point, increasing the likelihood of correct responses. Since deep learning systems process an enormous
18 volume of data thus creating knowledge and perform many complicated mathematical calculations, they require powerful hardware.
2 DEEP LEARNING ALGORITHMS
In spite of the fact that profound deep learning algorithms and calculations utilize self- learning representations, they depend on counterfeit neural systems (ANNs) that reflect the way the brain computes data. Calculations utilize obscure components within the input dispersion to extricate traits, gather objects, and find valuable information designs amid the preparing handle. This happens at diverse stages, using the calculation, a bit like educating machines for self-learning. Two sorts of learning are for the most part utilized in numerous applications: directed that is supervised or unsupervised. In the supervised learning, the algorithm looks at the preparing information and makes an evaluated work that can be utilized on modern occurrences. Unsupervised algorithms extricate experiences directly from the information itself, outlining or gathering it so that ready to make data-driven choices based on those insights.
In this computerized age, colossal sums of information that is, huge volume or amount of data is produced ordinary from social media, IoT and cloud computing advances etc. In Advanced learning strategies, it is conceivable to find representations that are utilized to extricate thoughts from crude truths automatically[1]. Profound learning calculations are an illustration of representation of learning strategies that utilize a progressive arrangement of nonlinear capacities to change over crude input information into more modern highlights that empower the acknowledgment of novel designs [2]. Deep learning models make utilize of different calculations. To select an calculation that can perform best on a particular assignment, it is critical to have a careful information of all Profound learning calculations.
Figure 1 Taxonomy of the deep learning algorithm
Deep learning algorithms, their characteristics and applications are summarized in table 1.
19 Table I. Deep learning Characteristics and applications
Algorithm Learning Behaviour Applications
Convolutional Neural
Networks (CNNs) Supervised It is useful for high- dimensional data, Keeps spatial information
Process medical images, predict time series and identify anomalies in satellite images
Long Short Term Memory
Networks (LSTMs) Supervised Long-term dependencies are
learned and memorized. Music composition and speech recognition
Recurrent Neural Networks (RNNs)
Supervised Stores temporal information and Captures time dependencies of data
Time-series analysis, machine translation, Image captioning, natural-language processing and handwriting recognition Generative Adversarial
Networks (GANs) Unsupervised Make new data instances that
look like the training data Develop realistic and animated figures,
Autoencoder Unsupervised inputs size are reduced to
make a smaller representation Image scanning, pharmaceutical invention, and popularity forecast Deep Belief Networks
(DBNs) Unsupervised Without a lot of
preprocessing, it's possible to get appealing results from raw data
Image and video recognition and motion-capture data
3 DEEP LEARNING IN HEALTHCARE
Deep Learning has been utilized in healthcare frameworks in later a long time. It is helping various restorative experts and examiners in picking up important information from clinical information and progressing restorative offices in distant better; a much better; a higher; a stronger; an improved">a much better way. A noteworthy inquire about field is evaluating patients' hazard of foreseeing a illness and conveying customized healthcare. By implies of a restorative imaging strategy, it is presently commonly utilized for sedate disclosure and distinguishing proof of life-threatening disorders such as diabetic retinopathy and cancer.
There could be a part of inquire about work that has as of now been done utilizing profound learning algorithms to anticipate the infection.. In this segment, the algorithms and their precision for conclusion is talked about. It'll be valuable for the analysts to recognize the most excellent calculation effortlessly for the forecast and conclusion of illness.
4 LITERATURE REVIEW
Within the later year, Farman Ali et al. [3] have determined a strategy that predicts heart infection some time recently a stroke or heart assault can happen utilizing back proliferation procedures and angle calculations. The exactness of proposed framework (98.5%) is compared with conventional classifier models and demonstrated that it is higher than the existing framework [13]. In another Heart infection recognizable proof system[4], the arbitrary woodland classifier calculation is the ideal machine learning calculation which gives the most elevated precision at 94.9% among other classification calculations specifically SVM , DT, RF and LR to foresee the heart disease.To choose basic highlights from the dataset, the univariate highlight choice and Help are utilized [14-15]. Tülay Karayılan; Özkan Kılıç[5],, have created a framework that can anticipate heart malady utilizing manufactured neural arrange- backpropagation procedure. Thirteen clinical highlights were utilized as input and the yield is confirmed with the exactness of 95%. In the survey of diabetic Within the survey of diabetic infection expectation, Motiur Rahman et al. [6] have created a navel diabetes classifying show based on Convolutional Long Short- term Memory (Conv-LSTM). When compared to other prevalent models such as CNN, Conventional LSTM and CNN-LSTM[16-18], the proposed demonstrate have gotten the greatest exactness of 97.26 percent over the Pima Indians Diabetes Database (PIDD).
Swapna G et al [7] have created an computerized diabetes discovery framework that utilized profound learning systems such as CNN and CNN-LSTM and utilized 5 fold cross approval to realize the most elevated precision of 95.1 percent among the robotized diabetes discovery strategies right now accessible. With forecast precision of 98.3 percent and 97.11 percent, Safial Islam Ayon et al.[8] created promising frameworks for the expectation of diabetes utilizing profound neural systems by preparing its traits in a five and ten fold cross-validation technique.
20 To get a more precise result, the proposed study employs two individual CNN architectures. to segment the Optic Cup (OC) and Optic Disc (OD).
This demonstrate is prepared and approved on the openly open DRISHTI–GS database, and it accomplishes a division precision of 98 percent for the optic cup and 97 percent for the optic disc.
H.N. Veena et al. [9] have planned a novel optic circle and optic container division procedure to analyze glaucoma. Within the proposed framework two diverse CNN structures, to section the Optic Glass (OC) and Optic Disc (OD), were utilized to urge more precise comes about. This show is prepared and tried on the freely open DRISHTI – GS database, and it accomplishes a division exactness of 98 percent for the optic plate and 97 percent for the optic glass. U Raghavendraa et al.[10] have proposed a modern CAD apparatus that employments profound learning methods to precisely distinguish glaucoma.
An 18 layer Convolutional Neural Arrange (CNN) was effectively prepared to extricate vigorous highlights from advanced fundus pictures, with a 98.13 percent precision.
N.Bhaskar et al.[11] have created a 98.6% exact programmed detection method for detecting chronic kidney disease using a help vector machine classifier and the Correlation Neural Network (CorrNN). Marjolein A. Heuvelmans et al.[12] developed a Lung Cancer Prediction Convolutional Neural Network (LCP-CNN) that can classify benign lung nodules with a 94.5 percent accuracy when trained on US screening data.
5 CONCLUSION
Deep Learning is a subset of Machine Learning that is used to solve difficult problems and construct intelligent solutions. To process the data and make predictions, deep learning employs Artificial Neural Networks. The techniques of Deep learning are being used by many medical professionals and academics to extract valuable knowledge from healthcare data and improve society's medical services. Deep Learning has made computer-aided disease prediction and computer-aided diagnosis possible in the healthcare field. In this paper, based on the previous research work done, the deep learning algorithms and their accuracy for predicting diseases is discussed and summarized. It will be useful for the researchers and medical experts to easily identify the best algorithm for disease prediction in the future.
Table 2 Summary of Disease Detection or Prediction
Ref. Year Cases Algorithm Accuracy
3 2020 Heart attack Prediction Feed-forward network - utilizes back propagation techniques and gradient
98.5%
4 2020 Heart disease
identification SVM, DT, RF and LR 94.9%.
5 2017 Heart disease prediction Artificial neural network
back propagation 95%.
6 2020 Diabetic Detection Convolutional Neural Network (CNN), Traditional LSTM (T-LSTM), and CNN-LSTM
97.26 %
7 2018 Diabetic Detection CNN and CNN-LSTM - 5 FC
layers 95.1%
8 2019 Diabetes Prediction Deep neural network and five-fold and ten-fold cross validation fashion
98.35%, - 5 FC 98.80% - 10 FC 9 2021 Diagnose glaucoma Deep learning
Convolutional neural network
98% - optic disc
97% - optic cup seg.
21 10 2018 Diagnosis of
Glaucoma Eighteen layer convolutional
neural networks (CNN) 98.13%
11 2020 Prediction of Chronic
Kidney Disease Support Vector Machine (SVM) classifier is
integrated with the CorrNN
98.67%.
12 2021 Lung cancer prediction Convolutional Neural
Network (LCP-CNN) 94.5 %
REFERENCES
1. [Bengio Y, Courville A, Vincent P. Representation learning: a review and new perspectives. IEEE Trans Pattern Anal Mach Intell. 2013;35(8):1798–828..
2. LeCun Y, Bengio Y, Hinton G. Deep learning. Nature 2015;521(7553):436–44.
3. Farman Ali, Shaker El-Sappagh, S.M. Riazul Islam, Daehan Kwak, Amjad Ali et al., “A Smart Healthcare Monitoring System for Heart Disease Prediction Based On Ensemble Deep Learning and Feature Fusion”,Information fusion, Volume 63, November 2020, Pages 208-222
4. Hager Ahmed a, Eman M.G. Younis a, Abdeltawab Hendawi b,c, Abdelmgeid A. Ali, “Heart disease identification from patients’ social posts, machine learning solution on Spark”, Future Generation Computer Systems, Volume 111, October 2020, Pages 714-722
5. Tülay Karayılan; Özkan Kılıç,” Prediction of heart disease using neural network”, 2017 International Conference.
6. Motiur Rahman, Dilshad Islam, Rokeya Jahan Mukti, Indrajit Saha, “A deep learning approach based on convolutional LSTM for detecting diabetes”, Computational Biology and Chemistry, Volume 88, October 2020
7. Swapna G, Soman KP, Vinayakumar R , “Automated detection of diabetes using CNN and CNN LSTM network and heart rate signals” Procedia Computer Science, Volume 132, 2018, Pages 1253- 1262 8. Safial Islam Ayon, Md. Milon Islam, “Diabetes Prediction: A Deep Learning Approach”, I.J. Information
Engineering and Electronic Business, 2019, 2, 21-27
9. H.N. Veena, A. Muruganandham, T. Senthil Kumaran, “A novel optic disc and optic cup segmentation technique to diagnose glaucoma using deep learning Convolutional neural network over retinal fundus images” Journal of King Saud University - Computer and Information Sciences , Available online 18 February 2021, in press
10. U Raghavendraa, Hamido Fujitab*, Sulatha V Bhandaryc, Anjan Gudigara, Jen Hong Tand, U Rajendra Acharyad,” Deep Convolution Neural Network for Accurate Diagnosis of Glaucoma Using Digital Fundus Images “,Information Sciences Volume 441, May 2018, Pages 41-49
