**3. Artificial intelligence (AI) applications in healthcare system**

In spite of few limitations, artificial intelligence (AI) are applied in healthcare system. Researchers mainly focus on the region of major three diseases: cardiovascular disease, nervous system disease, and life-threatening cancer also. In cardiology, [26] explained the prospective uses of the AI system for making a diagnosis of the cardiac diseases with the help of cardiac images. Cardiac stroke is a natural and commonly stirring disease that has an effect on more than 500 million people all around the world. It is the most leading cause of death in world. It has also high medical expenses across the world nearly about US\$ 689 billion, which causes serious trouble to patient families [27, 28]. For that reason, research on anticipation and medical treatment for stroke has a great impact. Recently, artificial intelligence (AI) processes have been used in additional and supplementary stroke-connected studies. In stroke-concerned cases, AI procedures help in the three main areas: before time for disease calculation and analysis, healing, and in addition to conclusion forecast and diagnosis assessment. About 85% of the time, stroke is caused by cerebral infarction, that is, thrombus in the vessel. For require of finding pre stroke indication, only some patients could obtain appropriate treatment. A movement-detecting device was developed for predicting early stroke [29]. For

**129**

*Application of Artificial Intelligence in Modern Healthcare System*

model structure resolution, two machine learning algorithms like PCA and genetic fuzzy finite state machine are mainly used. The revealing method is attached with a patient human action detection phase and the starting of the stroke detection phase. Ideally, the typical model is remarkably different from the patient movement, and an attentive model that can detect stroke can stimulate and assess medical action and make it immediately feasible. Correspondingly, a device that is wearable was proposed for gathering data for regular and pathological steps for calculation of stroke [30]. The data can be removed and copied by SVM and unseen Markov models, and this algorithm could suitably organize 91% of information to the exact group. For some identification of the stroke, neuro-imaging processes like CT scan and MRI are also essential for disease estimation. Several studies have attempted to concern machine learning techniques to neuro-imaging data to support with stroke analysis. SVM was used in resting-state functional MRI data, where endophenotypes of motor disability behind stroke were classified and recognized [31]. This algorithm can precisely distinguish patients with a precision of 87.6%. T1-weighted MRI, [32] helps to rearrange the stroke injury. This effect is similar for human-proficient physical injury explanation. Kamnitsas et al. [33] attempted 3D CNN aimed at injury fragmentation in multisculpt brain MRI. It likewise used fully associated provisional casual field representation for ultimate postprocessing of the CNN's soft segmentation plots. With the help of Gaussian process regression method, stroke anatomical MRI images were analyzed,and also establish the vortex pattern performed well than injury load/area like the expecting elements [34]. Machine learning (ML) techniques are also useful to examine stroke patients with CT scans. A free-floating intraluminal thrombus can be created like injury post stroke, and this is complicated to discriminate by carotid sign in CT imaging. Three machine learning (ML) algorithms were used to categorize two quantitative types: shape analysis with linear classification analysis, SVM, and artificial neural network [35]. Machine learning is also used in expecting and evaluating the presentation for stroke cure. In a critical emergency phase determination, the result of intravenous thrombolysis (tPA) has a sturdy link for the diagnosis per durance rate. With CT scan, SVM can be used for expecting whether the patients by thrombolysis (tPA) cure can build up suggestive intracranial hemorrhage [36]. In SVM, complete brain images were used as input, which acted healthier than traditional radiology-based procedures. For improving the medical result making procedure of thrombolysis (tPA) healing, a stroke treatment model was proposed for investigating perform guiding principle, clinical trials and meta-analysis with Bayesian principle network [37]. The model consisted of 56 different types of variables and 3 decisions aimed at investigating the process for analysis, cure, and effective calculation. An interaction tree was used, where the subgroup investigated suitable thrombolysis (tPA) dosage as per patient individuality, taking into consideration the healing efficacy and the possibility of bleeding [38]. Several issues can influence stroke diagnosis and syndrome mortality. Evaluating with traditional methods, machine learning techniques have returns in progressing calculation activity. To enhance and maintain the medical assessment making procedure, a model was proposed for expecting a three-month healing outcome by examining the physiological considerations for the duration of

48 hours following stroke with logistic degeneration [39]. A database was observed with 107 patient's medical information through acute anterior stroke and also posterior stroke via intra-arterial therapy [18]. Here, the data were examined through SVM and artificial neural network and achieved calculation accurateness of more than 70%. Machine learning procedures was used to recognize the control effect in brain arterio-venous abnormality satisfied with endo-vascular embolization. Though typical degeneration analysis representation could only reach a 43%

*DOI: http://dx.doi.org/10.5772/intechopen.90454*

### *Application of Artificial Intelligence in Modern Healthcare System DOI: http://dx.doi.org/10.5772/intechopen.90454*

*Alginates - Recent Uses of This Natural Polymer*

disease, which reaches over 90% accurate [22].

**3. Artificial intelligence (AI) applications in healthcare system**

In spite of few limitations, artificial intelligence (AI) are applied in healthcare system. Researchers mainly focus on the region of major three diseases: cardiovascular disease, nervous system disease, and life-threatening cancer also. In cardiology, [26] explained the prospective uses of the AI system for making a diagnosis of the cardiac diseases with the help of cardiac images. Cardiac stroke is a natural and commonly stirring disease that has an effect on more than 500 million people all around the world. It is the most leading cause of death in world. It has also high medical expenses across the world nearly about US\$ 689 billion, which causes serious trouble to patient families [27, 28]. For that reason, research on anticipation and medical treatment for stroke has a great impact. Recently, artificial intelligence (AI) processes have been used in additional and supplementary stroke-connected studies. In stroke-concerned cases, AI procedures help in the three main areas: before time for disease calculation and analysis, healing, and in addition to conclusion forecast and diagnosis assessment. About 85% of the time, stroke is caused by cerebral infarction, that is, thrombus in the vessel. For require of finding pre stroke indication, only some patients could obtain appropriate treatment. A movement-detecting device was developed for predicting early stroke [29]. For

Genetic data and EP plus image are all machine-comprehensible, that is why the machine learning (ML) algorithms can be straightly presented after quality control processes or appropriate preprocessing. Though huge extents of medical data are like descriptive content, like a substantial examination, operative notes, and an experimental laboratory reports and release abstracts, these are formless and inconceivable for computer programming. Below this background, natural language processing (NLP) targets removing helpful data from the descriptive text to support the medical conclusion making [3]. A natural language processing (NLP) pipeline includes two main components: (i) classification and (ii) text processing. During text processing, the natural language processing (NLP) recognizes a sequence of disease-appropriate keywords at clinical remarks related to the past records [22]. After that, keyword subsets are preferred during analyzing their achievements in the arrangement in the normal abnormal cases. The authorized keywords then enter and enhance the controlled information to support medical choice making. The natural language processing pipelines have been developed to help the medical choice making on attentive treatment preparations and monitoring critical effects. For instance, it was showed that establishment of natural language processing, for analyzing the chest X-ray reports would help the antibiotic assistant system to aware physicians for the probable necessitate for anti-infective therapy [23]. Natural language processing was used to mechanically monitor laboratory-based difficult effects. Moreover, the natural language processing pipelines can also assist with disease analysis [24]. A recognized of 14 cerebral aneurysm disease-associated changeable during executing natural language processing (NLP), based on the clinical remarks [25]. Resulting variables are effectively applied for categorizing the common patients and the patients with cerebral problems, with 86% to 95% accuracy rates on the validation and training trials correspondingly. A natural language processing was implemented to extort the peripheral arterial diseaseallied keywords from description clinical remarks. The keywords are then applied to categorize the common patients and the patients who have peripheral arterial

**2.5 Natural language processing**

**128**

model structure resolution, two machine learning algorithms like PCA and genetic fuzzy finite state machine are mainly used. The revealing method is attached with a patient human action detection phase and the starting of the stroke detection phase. Ideally, the typical model is remarkably different from the patient movement, and an attentive model that can detect stroke can stimulate and assess medical action and make it immediately feasible. Correspondingly, a device that is wearable was proposed for gathering data for regular and pathological steps for calculation of stroke [30]. The data can be removed and copied by SVM and unseen Markov models, and this algorithm could suitably organize 91% of information to the exact group. For some identification of the stroke, neuro-imaging processes like CT scan and MRI are also essential for disease estimation. Several studies have attempted to concern machine learning techniques to neuro-imaging data to support with stroke analysis. SVM was used in resting-state functional MRI data, where endophenotypes of motor disability behind stroke were classified and recognized [31]. This algorithm can precisely distinguish patients with a precision of 87.6%. T1-weighted MRI, [32] helps to rearrange the stroke injury. This effect is similar for human-proficient physical injury explanation. Kamnitsas et al. [33] attempted 3D CNN aimed at injury fragmentation in multisculpt brain MRI. It likewise used fully associated provisional casual field representation for ultimate postprocessing of the CNN's soft segmentation plots. With the help of Gaussian process regression method, stroke anatomical MRI images were analyzed,and also establish the vortex pattern performed well than injury load/area like the expecting elements [34]. Machine learning (ML) techniques are also useful to examine stroke patients with CT scans. A free-floating intraluminal thrombus can be created like injury post stroke, and this is complicated to discriminate by carotid sign in CT imaging. Three machine learning (ML) algorithms were used to categorize two quantitative types: shape analysis with linear classification analysis, SVM, and artificial neural network [35]. Machine learning is also used in expecting and evaluating the presentation for stroke cure. In a critical emergency phase determination, the result of intravenous thrombolysis (tPA) has a sturdy link for the diagnosis per durance rate. With CT scan, SVM can be used for expecting whether the patients by thrombolysis (tPA) cure can build up suggestive intracranial hemorrhage [36]. In SVM, complete brain images were used as input, which acted healthier than traditional radiology-based procedures. For improving the medical result making procedure of thrombolysis (tPA) healing, a stroke treatment model was proposed for investigating perform guiding principle, clinical trials and meta-analysis with Bayesian principle network [37]. The model consisted of 56 different types of variables and 3 decisions aimed at investigating the process for analysis, cure, and effective calculation. An interaction tree was used, where the subgroup investigated suitable thrombolysis (tPA) dosage as per patient individuality, taking into consideration the healing efficacy and the possibility of bleeding [38]. Several issues can influence stroke diagnosis and syndrome mortality. Evaluating with traditional methods, machine learning techniques have returns in progressing calculation activity. To enhance and maintain the medical assessment making procedure, a model was proposed for expecting a three-month healing outcome by examining the physiological considerations for the duration of 48 hours following stroke with logistic degeneration [39]. A database was observed with 107 patient's medical information through acute anterior stroke and also posterior stroke via intra-arterial therapy [18]. Here, the data were examined through SVM and artificial neural network and achieved calculation accurateness of more than 70%. Machine learning procedures was used to recognize the control effect in brain arterio-venous abnormality satisfied with endo-vascular embolization. Though typical degeneration analysis representation could only reach a 43%

precision rate, this technique's exertion is much enhanced with 97.5% exactness. An optimal algorithm was analyzed to calculate 30 days mortality test and gained additional exact calculation than surviving techniques [40]. Likewise, SVM was used to calculate the stroke mortality via discharge. Additionally, the application of the synthetic alternative oversampling procedure was proposed to decrease the stroke effect calculation prejudice reasoned among class inequality between several datasets. Brain images were examined for calculating the effect of stroke cure. CT scan data were examined through machine learning procedure for estimating the cerebral edema through hemispheric infraction [41]. A random forest was constructed to involuntarily recognize the cerebrospinal fluid (CSF) and examined the changes in the CT scan, and this is more precise and capable compared to the traditional procedures. Functional connectivity was extracted from magnetic resonance imaging (MRI) and practical magnetic resonance imaging (MRI) data, and ridge degeneration and multitasking intellect were also applied for cognitive deficit calculation following stroke [42]. A relationship was examined, which involved injuries extorted from magnetic resonance imaging (MRI) and the cure effect through Gaussian method regression technique [43]. The model was used to calculate the difficulty of cognitive damages during stroke and the way of retrieval in due course. In Arterys Cardio DL process, where artificial intelligence (AI) is help to make available programmed and also changeable ventricle segmentations related on traditional MRI of cardiac images [44]. In nervous system disease, an artificial intelligence (AI) method was developed [45] for repairing the regulation of body movement in quadriplegia patients. Farina et al. experienced the control of the offline man–machine edge, which applies the release timings for the spinal motor neurons for controlling the prosthesis of the upper limb. IBM Watson for the oncology diagnosis can be a consistent AI for cancer diagnosis from start to the end, which was explained by Somashekhar et al. [46] by a double-blinded validation study. A clinical image was examined for recognizing skin cancer subtypes [20]. The applications of these three types' diseases are not absolutely unpredicted. These three diseases are principal death causes; for that reason, analyzing the stages of the disease before time is vital to avoid worsening of the patients' health condition. Moreover, quick diagnoses can prospectively reach throughout recovering the analysis measures on electrophysiological (EP) or electronic medical record (EMR), imaging and genetic, and this is the major power of the artificial intelligence (AI) technique. Moreover, apart from the three main diseases, artificial intelligence (AI) system has been used in another disease too: to examine the ocular image data for diagnosing inherited cataract diseases [19]. A referable diabetic retinopathy was detected by the retinal fundus photographs [21].
